Three Essays on European Union Advances Toward a Single Currency and Its Implications for Business and Investors Charlotte Anne Bond Old Dominion University

Home , European Currency Unit, Greek drachma, Portuguese escudo, Spanish peseta, World currency

Old Dominion University ODU Digital Commons Theses and Dissertations in Business College of Business (Strome) Administration

Winter 1998 Three Essays on European Union Advances Toward a Single Currency and Its Implications for Business and Investors Charlotte Anne Bond Old Dominion University

Follow this and additional works at: https://digitalcommons.odu.edu/businessadministration_etds Part of the Finance and Financial Management Commons, International Business Commons, and the International Relations Commons

Recommended Citation Bond, Charlotte A.. "Three Essays on European Union Advances Toward a Single Currency and Its Implications for Business and Investors" (1998). Doctor of Philosophy (PhD), dissertation, , Old Dominion University, DOI: 10.25777/mc19-6f14 https://digitalcommons.odu.edu/businessadministration_etds/77

This Dissertation is brought to you for free and open access by the College of Business (Strome) at ODU Digital Commons. It has been accepted for inclusion in Theses and Dissertations in Business Administration by an authorized administrator of ODU Digital Commons. For more information, please contact [email protected]. Three Essays on European Union Advances Toward A Single Currency and its Implications for Business and Investors

Charlotte Anne Bond

A dissertation submitted to the Faculty of Old Dominion University in partial fulfillment of the requirements for the degree of

Doctor o f Philosophy

(Finance)

Old Dominion University College of Business Norfolk, Virginia (December 1998)

Approved by:

Mohammad Najand (Committee Chair)

\ tee Member)

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number: 9921767

This microform edition is protected against unauthorized copying under Title 17, United States Code.

UMI 300 North Zeeb Road Ann Arbor, MI 48103

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ABSTRACT

Three Essays on European Union Advances toward a Single Currency and its Implications for Business and Investors

Charlotte Anne Bond Department of Business Administration Old Dominion University Committee Chair: Mohammad Najand

The first chapter examines the changes in various European currencies’ exchange rates through the time period 1980 through 1997. Specifically, we are interested to determine if there is any affect to the volatility of these exchange rates and specific events related to the advancement of European Unification. In order to move to a single currency it is imperative that the separate currencies become less volatile to facilitate the move to a single currency. In this study, we examine whether this is the case and discuss which currencies appear to display this behavior. It is observed that of the 14 currencies examined all but Ireland and Italy’s currencies see dramatic reductions in volatility. The second chapter examines the effects of announcements concerning European Monetary Union on the exchange rate volatilities of several European currencies. It is expected that when good news is portrayed in regard to a single currency, this will be considered bad news, thus eliciting a negative reaction. The currencies examined are the German mark, the Portuguese escudo, the Italian lira, the Greek drachma, and the Spanish peseta. In terms of volatility, a reaction to good news should be a reduction in volatility, as bad news should cause an increase in volatility. In total there are 22 announcements examined from January 1990 through September 1997. The German mark is observed to experience greater increases in volatility than decreases as does the Italian lira. Portugal and Greece appear to react more strongly to positive news in that the decreases in volatility are on average greater than the increases. In the third chapter, the reactions of volatility changes to the returns of American Depository Receipts of companies from European Union member nations are examined. It is examined whether announcements regarding European Monetary Union create a notable change in the volatility of returns of these instruments. If a single currency is viewed as good news for these companies, the volatility of the returns of these companies should decrease. If the advent of a single currency is bad news, the volatility of returns should increase. In total there are 10 announcements examined from January 1990 through September 1997. Of the 8 countries examined, Finland, France and the Netherlands display no notable reactions. Luxembourg witnesses the largest decreases in volatility around 6 of the ten dates examined.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. To Byron and Austen

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Acknowledgements

Were it not for the assistance and patience of Mohammad Najand, this feat would have taken years if it had ever been completed. For this I am forever grateful. As well, I acknowledge the useful comments and moral support of Sylvia Hudgins and Vinod Agarwal. I thank you all.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TABLE OF CONTENTS

PAGE

DEDICATION ...... iii

ACKNOWLEDGEMENTS ...... iv

LIST OF TABLES ...... vii

LIST OF FIGURES ...... viii

CHAPTER

I. Changes in European Currency Volatility as Related to Changes Occurring during Europe 1992

A. Introduction ...... 1 B. Literature Review ...... 3 C. Data and Methodology i. Data ...... 8 ii. Methodology ...... 9 D. Empirical Results ...... 11 E. Conclusions ...... 17

II. Volatility Changes in European Currency Exchange Rates Due to EMS Announcements

A. Introduction ...... 18 B. Literature Review ...... 21 C. Data and Methodology i. Data ...... 25 ii. Methodology ...... 26 D. Empirical Results ...... 29 E. Conclusions ...... 33

III. Volatility Changes in European American Depository Receipt Returns Evidence from NASDAQ Market

A. Introduction ...... 35 B. Literature Review ...... 37

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. C. Data and Methodology i. Data ...... 38 ii. Methodology ...... 39 D. Empirical Results ...... 43 E. Conclusions ...... 47

REFERENCES ...... 48

CURRICULUM VITA ...... I l l

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF FIGURES FIGURE PAGE

1-1 Volatility: Austrian schilling ...... 53

1-2 Volatility: Belgian franc ...... 54

1-3 Volatility: Danish kroner ...... 55

1-4 Volatility: Finnish markka ...... 56

1-5 Volatility: French franc ...... 57

1-6 Volatility: German mark ...... 58

1-7 Volatility: Greek drachm a ...... 59

1-8 Volatility: Irish pound...... 60

1-9 Volatility: Italian lira...... 61

1-10 Volatility: Dutch guilder ...... 62

1-11 Volatility: Portuguese escudo...... 63

1-12 Volatility: Spanish peseta ...... 64

1-13 Volatility: Swedish kronor ...... 65

1-14 Volatility: British pound ...... 66

1-15 Percent Change in Exchange Rate Volatility from 1979-1985 to 1986-1992 ...... 67

1-16 Percent Change in Exchange Rate Volatility from 1986-1992 to 1993-1998 ...... 68

1-17 Percent Change in Exchange Rate Volatility from 1979-1985 to 1993-1998 ...... 69

2-1 Volatility: German mark ...... 70

2-2 Volatility: Portuguese escudo...... 71

2-3 Volatility: Italian lira ...... 72

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. FIGURE PAGE

2-4 Volatility: Greek drachma ...... 73

2-5 Volatility: Spanish peseta ...... 74

2-6 Percent Change per Event: Germany ...... 75

2-7 Percent Change per Event: Portugal ...... 76

2-8 Percent Change per Event: Italy ...... 77

2-9 Percent Change per Event: Greece ...... 78

2-10 Percent Change per Event: Spain ...... 79

3-1 Volatility: Finnish ADRs ...... 80

3-2 Volatility: French ADRs ...... 81

3-3 Volatility: Greek ADRs ...... 82

3-4 Volatility: Irish ADRs ...... 83

3-5 Volatility: Luxembourg ADRs ...... 84

3-6 Volatility: Dutch ADRs ...... 85

3-7 Volatility: Swedish ADRs ...... 86

3-8 Volatility: British ADRs ...... 87

3-9 Percent Change per Event: Finland ...... 88

3-10 Percent Change per Event: France ...... 89

3-11 Percent Change per Event: Greece ...... 90

3-12 Percent Change per Event: Ireland ...... 91

3-13 Percent Change per Event: Luxembourg ...... 92

3-14 Percent Change per Event: Netherlands ...... 93

3-15 Percent Change per Event: Sweden ...... 94

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. FIGURE PAGE

3-16 Percent Change per Event: U.K...... 95

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF TABLES

TABLE PAGE

1-1, Panel A GARCH Estimates for the Full Period (January 1979 - April 1998)...... 96

1-1, Panel B Diagnostics of Full Time Period...... 97

1-2 GARCH Estimates for the First Subperiod (January 1979 - December 1985) ...... 98

1-3 GARCH Estimates for the Second Subperiod (January 1986 - December 1992) ...... 99

1-4 GARCH Estimates for the Third Subperiod (January 1993- April 1998) ...... 100

1-5 Averages of Daily Volatility per Subperiod and Percent Changes In Volatility between Subperiods ...... 101

2-1 Summary of Announcements Obtained from theWall Street Journal 102

2-2, Panel A AR( 1) - EGARCH (1,1) Estimates for Period (January 1979 - April 1998)...... 103

2-2, Panel B Diagnostics of AR(1) - EGARCH(1, 1) Results ...... 103

2-3 Percent Changes in Average Daily Volatility One Month prior to and after Announcements ...... 104

3-1 Summary of Announcements Obtained from theWall Street Journal 106

3-2, Panel A GARCH Estimates for the Country ADR portfolios...... 107

3-2, Panel B Diagnostics of the ADR portfolios ...... 108

3-3 Percent Changes in Average Daily Volatility One Month prior to and after Announcements ...... 109

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1

Chapter I

Changes in European Currency Volatility as Related to Changes Occurring during

Europe 1992

A. Introduction

In this study we are interested in the development of a single currency in Europe,

now known as the Euro, as an international or world currency. In his seminal work,

Mundell (1961) characterizes an optimum currency area as a region within which there is

factor mobility but has factor immobility with all areas outside this region. The

development of the European Union (EU) over the last years will certainly support the

former aspect of this statement. However, if the Euro is supported by an optimum

currency area, as might be the case, we are interested in it as a world currency and more

specifically the characteristics of its development as such.

As international money, a currency should be a reliable store of value. The ECU

(as a basket of currencies) has been the world’s third most important currency for

denomination of long-term loans after the US dollar and the German mark. For holders

of European currencies the ECU has been seen to be a better store of value than either the

US dollar or special drawing rights (SDR) (Pozo, 1987). This is determined by finding

that the average monthly exchange rates of European currencies to the ECU is less

variable than the comparable rates to the US dollar or the SDR.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Mundell (1961) suggests that the level of capital mobility is the key determining

factor of an optimum currency area. It will be interesting to see which European

currencies are becoming less volatile in order to facilitate the move to a single currency.

Argimon and Roldan (1994) find high capital mobility between the Netherlands,

Germany, and the United Kingdom. They also find low capital mobility between Spain,

France, Italy, Denmark, Belgium, and Ireland. Similarly, Helg, Manasse, Monacelli, and

Rovelli (1995) find the “perific” countries of Greece, Ireland, Spain, and Portugal to have

low levels of specialization and low levels of correlation of industries within the country

with regard to growth. This can easily be interpreted as a symptom of low factor

mobility. As factors become more mobile, specialization will take place in countries that

dominate performance in that industry. Low factor mobility not only hurts a regional

bloc member’s integration within its bloc, it also will be detrimental to the strength of the

member’s economy.

As the EU makes plans to change over to a single currency, one becomes curious

as to whether this will be more beneficial economically than maintaining a target zone

currency regime. Poole (1970) suggests that the best exchange rate regime is the one

which delivers the lowest variance of some target variable, such as output or prices, given

the presence of exogenous stochastic shocks to the economic system. Target zones offer

more stability than either a fixed or flexible exchange rate regime as demonstrated by

Sutherland (1995). In that study, he finds the optimal bandwidth will depend on the

relative variance of the shocks and will increase as its contribution of velocity increases

relative to the demand shocks. This demonstrates that a target zone offers a compromise

between the ability of fixed exchange rates to deal with velocity shocks and the ability of

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. flexible exchange rates to deal with good demand shocks. However, a single currency

should eliminate much of this concern, thus dominating all three options of target zones,

fixed exchange, or flexible exchange rate systems.

B. Literature Review

Much effort has been devoted to modeling exchange rates in financial literature.

Among the many interests in this area, one of particular interest is to find the correct

specification of the monetary model or to make this elusive model work as theory

suggests. Meese and Rogoff (1983) (MR) determine that macroeconomic theory does not

adequately explain exchange rate changes. Schinasi and Swamy (1989), in contrast, use

variable coefficients rather than the fixed coefficients of MR. They find that depending

on the assumptions and the specific model one-step ahead and multi-step ahead models

with varying coefficients outperform the random walk model when forecasting exchange

rates thus finding support for the monetary model. Noting that Krugman (1991) and

Froot and Obstfeld (1991) find that exchange rates are both linearly and non-linearly

related to the fundamentals, Chinn (1991) uses a method he calls alternative conditional

expectations (ACE) to model exchange rates. He finds ACE provides superior in-sample

results but is sometimes outperformed by non-linear models out-of-sample. MacDonald

and Taylor (1994) suggest that it is the timing and dynamics of the model which are not

being considered correctly, rather than an inherent flaw in the monetary model. These

authors believe that research should take a long-run view rather than the typically taken

short-run view when testing this model. By using a multivariate cointegration technique,

the authors find significant cointegration between the spot exchange rate and the

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4

fundamentals that adequately forecast up to 24 months out-of-sample. Their model is

found to dominate the typically used first differences model which is seldom seen to

outperform a random walk.

Part of the problem with modeling exchange rates is the use of the official,

usually managed, exchange rates (Phylaktis and Katsimatis, 1994). When using the black

market rate, which is allowed to react naturally to actual and anticipated changes in

prices, and measuring their properties with seemingly unrelated regressions (SUR(E))

purchasing power parity (PPP) is found to be more likely the case and a fifty percent

correction in PPP after a shock would occur in approximately a year. This is in contrast

to Abuaf and Jorion (1990) who use generalized least squares (GLS) regressions and

determine it would take 3 to 5 years for PPP to obtain a fifty percent correction following

a shock. The real exchange rate long-term stability is a result of changes in prices due to

the volatile nature of nominal exchange rates and there is mean reversion in real

exchange rates according to Phylaktis and Katsimatis (1994) for the studied countries.

Unfortunately, we generally do not have access to the black market rates and have to

hope that the official rates are an adequate representation of what we are trying to

measure.

In more recent literature, the use of ARCH, GARCH and their variations have

become popular methods of modeling and measuring foreign exchange rates.

Autoregressive conditional heteroskedasticity (ARCH) models allow and measure the

changing variance of variables in a system. In financial studies, variance is of great

importance. According to the capital asset pricing model (CAPM), it is the variance

(risk) of a stock’s (or instrument's) return to the market’s return that determines its price

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5

(return). It is also the variance of a stock's price that determines the value o f its options

contracts. Similarly, many financial instruments’ values are at least in part a function of

their variance of return (or price). Given that this variance is assumed stationary by many

methods of measurement, particularly a standard univariate, bivariate, or multivariate

regression, these methods fail to adequately model variables and systems that have a

variance which is subject to change. For this very reason, ARCH and generalized ARCH

(GARCH) are an appropriate choice and have been used extensively in the literature to

model exchange rates.

Essentially developed in Engle (1982), ARCH models have been extended in

several ways to suit different purposes and fit different processes and systems. Important

works involving various ARCH and GARCH models to measure foreign exchange

include Baillie and Bollerslev (1989). Daily, weekly, bi-weekly, and monthly exchange

rates are examined. The daily series is seen to have a unit root and to be well represented

by a GARCH model. As the series is aggregated into less frequent measurements, the

series becomes more normal and is less well represented by either GARCH or ARCH.

ARCH models are also used to measure risk premia in the foreign exchange market by

modeling 30-day forward rates with spot rates in Baillie and Bollerslev (1990). In that

study the standard asset pricing model does not hold, but rather they find inefficiency in

the market such as significant first differences.

Specifically related to European Monetary System (EMS), Bollerslev (1990)

models the coherence and correlations of the exchange rates in the EMS period (post

March 1979) and compares it to the pre-EMS period (before March 1979) of the “snake”

system. Using weekly data, Bollerslev finds correlations to be higher post-March 1979

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6

for EMS and non-EMS countries. That study finds it difficult to reject a random walk,

but also finds little evidence against a GARCH (1,1) model.

Due to exchange rates’ changing volatility and stability and their leptokurtic

distributions, more traditional modeling techniques, specifically those which assume

constant variance, such as standard regression are not adequate models as noted in Mussa

(1979) and Friedman and Vandersteel (1982). In modeling various exchange rates with

respect to the U.S. dollar, Hseih (1988) finds the conditional distribution of daily

exchange rate returns to changethrough time and an ARCH (12) model does an adequate

job of capturing this. As frequency of observation decreases so does the adequacy of

ARCH models in modeling exchange rates as noted by Diebold (1988) and Baillie and

Bollerslev (1989), thus daily data is generally better represented than monthly data.

ARCH and GARCH models have been seen to be useful in measuring information

processing in foreign exchange markets. Specifically, Engle, Ito and Lin (1990) show

information processing is a source of volatility clustering such that each market’s

volatility is significantly affected by changes in another market’s volatility.

One problem noted is that GARCH models make it difficult to evaluate whether

shocks to variance persist. Nelson (1991) presents an exponential ARCH model which

has a linear process whose stationarity is easily checked. This method is used in cases

where shocks produce asymmetric results.

Integrated GARCH, I-GARCH, is a class of models which are integrated in

variance as discussed in Engle and Bollerslev (1986). This is useful for measuring

persistence. In foreign exchange, IGARCH is often used to determine the persistence of

volatility shocks. Integration in variance is identified by the sum of the coefficients of a

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7

model to be equal to or very close to one. In Engle and Bollerslev (1986) the coefficients

sum to 0.996. Other studies including Bollerslev (1987), Hseih (1988), Baillie and

Bollerslev (1989), Taylor (1990) and many others have similar findings. As to whether

there is co-persistence among the variances is examined by Bollerslev and Engle (1990)

which finds evidence to suggest a set of underlying forcing variables using bivariate

GARCH (1, 1). This evidence could be of great importance for further modeling of

portfolio allocation.

Another related model is ARCH in Mean (ARCH-M) from Engle, Lilien and

Robins (1987) in which the mean is conditional and a function of the variance such that

an increase in the variance will find either an increase or a decrease in the conditional

mean. This model is useful when studying the mean-variance trade off situations which

are very common in financial research. For a fairly comprehensive discussion on the use

of ARCH and GARCH along with their variations, Bollerslev, Chou, and Kroner (1992)

have prepared a summary.

Exponential GARCH (EGARCH), as developed in Nelson (1991) is seen to

provide an adequate representation of the volatility found in EMS countries’ currencies

exchange rates (Hu, Jiang, and Tsoukalas, 1997). Due to the arrangements inherent in

EMS, there may be asymmetry between countries’ reactions to volatility shocks.

EGARCH provides a model specification which allows separate effects of good and bad

news along with a structure to examine persistence of the volatility.

The Hu, Jiang, and Tsoukalas (1997) study is similar to the one proposed here.

However, that study’s (1) data set ends before 1992 so that it cannot encompass the

events studied here, (2) they use weekly data whereas this study examines daily exchange

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. rates. (3) Their study uses rates in relation to the German mark, whereas in this study US

dollar rates are utilized and (4) we use a larger sample of member countries, they only

examine the original 12 member states. Finally, (5) this study uses the return of the ECU

as an independent variable in the model. This is seen to improve the model’s results

substantially.

In this study we propose to examine changes in the volatility of 14 European

countries’ currency exchange rates per US dollar as Europe changed with progress

toward a single economy. We hypothesize that as Europe experiences important events

toward its development as an integrated economic bloc the individual currencies of the

affected nations will become more stable as witnessed through decreased volatility in

their exchange rates. The events considered here are (1) the declaration of a program

which became known as “Europe 1992” in 1986 and (2) the time at which this program

was scheduled to be completed in December 1992.

C. Data and Methodology

C-i. Data

The data used in this study are the daily exchange rates of several European

currencies relative to the U.S. dollar. Austrian schilling, Belgian franc, Danish kroner,

Finnish markka, French franc, German mark, Greek drachma, Irish pound, Italian lira,

Portuguese escudo, Spanish peseta, Swedish kroner, UK pound, and ECU per US dollar

rates are used in this study. This data is obtained from the United States Federal Reserve

Bank of New York. Three individual periods will be examined. The first period is 1979

to 1985, which is the time prior to the proposal of a single Europe by Jacque Delors in

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9

Europe 1992. The second period is 1986 to 1992, which is the development period

during which Europe prepared itself for all of the changes scheduled to begin no later

than January 1, 1993. The final period will be 1993 to April 1998, the period after trade

barriers were to be removed.

C-ii. Methodology

AR(1) - EGARCH (1,1) models are used for the currencies to measure the daily

volatility. Engle introduced the autoregressive conditional Heteroskedasticity (ARCH)

model in Engle (1982). This model allows the conditional variance to change over time

as a function of past errors. The strength of this model is that the conditional means and

variances can be estimated jointly using traditional specified models for economic

variables.

In this model, Yt is a random variable whose mean is given by Xtp (independent

variables) and is a linear combination of lagged endogenous and exogenous variables

included in the information set Ot-i with p, a vector of unknown parameters.

Yt |

ht = oc0 + £iCtiee2t-i (1)

e,= =Y,-X,P

Bollerslev (1986) extends the ARCH process to GARCH (Generalized

Autoregressive Conditional Heteroskedastic), which allows for a more flexible lag

structure. Bollerslev points out that the extension of the ARCH process is very much like

the extension of the standard time series process to the general ARMA process.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 10

The GARCH (p, q) regression model is obtained by

6, = Yt - X,p

st-i I

h, = a 0 + Zi=iqaiS2t.i + Zj=ipPih,_j

p > 0 q > 0

Where ao> 0 a, > 0 i = q-

Pi > 0 i = l,...,p.

For p = 0, the process reduces to the ARCH (q) process, and for p = q = 0, Sj is just white

noise. Bollerslev shows that the resulting GARCH (p, q) model is essentially a stationary

ARCH(q) process. We utilize the following GARCH (1,1) model to study the impact of

these specific announcements on the exchange rate volatility.

Rt = Po + PiRt-i + P2 R-ECUt + st

St., | O,., N(0, ht) (3)

ht = ao + a ,h t-i + a2S2t-i

Where Rt is defined as log (St/ St-i) * 100, where St is the spot exchange rate at time t (as

in Baillie and Bollerslev, 1989), R-ECUt as log ((ECU/USSV (ECU/US$) m ) * 100, and

ht is variance of et and is calculated recursively by a system of equations (3).

Bollerslev shows that in a GARCH (p, q) process the orders of p and q can be

identified by applying the traditional Box and Jenkins time series techniques to the

autocorrelations and partial autocorrelations for the squared process of et. Since the

autocorrelation and partial autocorrelation for the squared residuals from model (3) cut

off after lag one, we selected GARCH (1,1) as the appropriate model. Bollerslev (1986)

also shows that GARCH (1,1) adequately fits many economic times series.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 11

A limitation of the GARCH model described above is that the conditional

variance responds to positive and negative residuals, st-i, in the same manner. However,

empirical evidence in financial time-series shows that there is a negative correlation

between the current returns and future return volatility. The GARCH model imposes the

nonnegative constraints on the parameters, ai and yi, while there are no restrictions on

these parameters in an exponential GARCH (EGARCH) model proposed by Nelson

(1991). In the EGARCH (1, 1) model, the conditional variance, ht, is an asymmetric

function of lagged residuals et-i:

Rt = Po + PiRt-i + P2 R-ECU, + e, (4)

ln(ht) = co + a , g (zt- i ) + yi ln(hn)

where g(zt) = 0zt + y[|zt| - E|zt|] and zt = et/Vht. Consider the g(zt) function above. If zt is

positive then g(zt) is a linear function of the slope changes, zt, with slope (0 + y). If zt is

negative then the slope changes to (0 - y). Consequently, the conditional variance ht

responds asymmetrically to the sign of innovation zt.i.

D. Empirical Results

The estimates of the AR(1)-EGARCH model for the full period and the three

subperiods are given in Tables 1-1 through 1-4. These tables include the coefficients for

the return on the ECU, the lag of the return of the respective currency, the exponential

ARCH (ai) component, the exponential GARCH (yi) component, and the theta (0)

component. Several interesting findings are seen in the estimates of the full period. First,

we observe that a AR(1) - EGARCH (1,1) model generally fits very well. This is

demonstrated both in the highly significant coefficients for each country and the high R2.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 12

For Austria the return on ECU, the lag component of the return on the schilling, the ai

coefficient, and the yi coefficient are all significant at the one-percent level of

significance. The 0 coefficient is not significant at any conventional level in this case.

Belgium shows similar results for the full period except that the 0 coefficient is

significant at the one-percent level. Denmark again provides similar results in that all of

the relevant coefficients are significant at the one-percent level with the exclusion of its 0

coefficient, which is not significant at any conventional level. The rest of the table shows

very similar results including that only one other country does not have a significant 0

coefficient with that country being France. In Panel B, we report the diagnostics for the

EGARCH (1,1) model. The Akaike Information Criteria (AIC) and the Log Likelihood

(LnL) are used to measure the appropriateness of the model for the given data. Also, for

non-linear time series models, the portmanteau Q-test statistics (Q) based on standardized

residuals (st/Vht) is used to test for non-linear effects. The Q (10) statistic cannot reject

the null hypothesis of no nonlinear effects for up to lag 10 for any of the 14 currencies.

Thus it appears that the nonlinearity in the volatility series has been successfully removed

by our GARCH model specifications. Also reported is the LaGrange multiplier test (LM)

for ARCH disturbances proposed by Engle (1982) in Panel B. The null hypothesis that

the disturbances lack ARCH effects is not rejected.

(Insert Table 1-1 here)

The measures of volatility may be observed graphically in Figures 1 through 14.

As one might notice some of these charts display obvious reductions in volatility as time

progresses. Belgium, Denmark, France, and Germany are somewhat obvious in this

respect. Others, especially those such as Austria, Finland, Greece, Ireland, Portugal,

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 13

Spain, and Sweden that experience such high volatility in a few isolated incidents that the

charts are hard to interpret, whereas the rest are just ambiguous.

(Insert Figures 1-1 through 1-14 here)

The estimates for the first subperiod (1979 through 1985) are presented in Table

1 -2. Again, for every country all relevant coefficients (the return on the ECU, the lag of

the return of the respective currency, the exponential ARCH (cxi) coefficient, the

exponential GARCH (yi) coefficient, and the theta (0) coefficient) are significant at least

at the 5 percent level, but the lion’s share are significant at the one percent level.

Interestingly, all 0 coefficients are highly significant in this period with the one exclusion

of Ireland. This includes the three countries’ (Austria, Denmark, and France) whose 0

coefficients are not significant in the full period.

(Insert Table 1-2 here)

For the second subperiod (1986 through 1992), as presented in Table 1-3, the

results are similar with all relevant coefficients being statistically significant with the two

exclusions of the theta coefficients for Denmark and Spain.

(Insert Table 1-3 here)

The results for the third subperiod, as presented in Table 1-4, are again quite

similar with only three countries, Denmark, Finland, and UK not having significant 0

coefficients, while all other relevant coefficients are highly statistically significant.

(Insert Table 1-4 here)

Initially, the results found for the full period and its three subperiods suggest that

the AR(1) - EGARCH(1, 1) model fits very well, but this also demonstrates a few other

interesting points. The return on the ECU is a very important factor in this model. In the

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 14

full period this coefficient ranges from 0.67 for Sweden to 1.01 for Germany, all

significant at the one-percent level (see Table 1-1). This suggests that each currency is

heavily influenced by movement in the ECU with some countries (Germany, Belgium,

Austria, and France) moving almost exactly in tandem with the ECU, given that nothing

else is changing.

Table 1-2 presents the results for the first subperiod. The coefficients for the

ECU are again quite interesting, ranging from 0.56 for UK to 0.92 for Austria. The

coefficients are generally not as close to one as in the full period. This demonstrates that

early in the development of the ECU the individual currencies are not as closely tied, but

still quite impressively tied.

In Table 1-3, this influence is observed to increase in the second subperiod as

witnessed in the ECU’s coefficients ranging from 0.72 for Finland to 1.01 for Germany

with all but two of these coefficients (Finland and Sweden) being 0.80 or greater. In the

third subperiod (Table 1-4), the coefficients are still highly significant, but the magnitude

is generally greater. In this period the coefficients range from 0.63 for UK to 1.10 for

Austria with 5 (Portugal, Germany, Netherlands, Belgium, and Austria) being greater

than 1.00 and 5 others (Greece, Spain, Finland, Denmark, and France) being greater than

0.95. This implies that not only are the individual currencies moving in tandem with the

movements of the ECU, some are actually overshooting that movement even if to a very

small extent.

The lag coefficient of the model is generally significant in all periods for all

countries with the exception of Sweden and Italy in the third period. However, its

influence is not as great as that of the ECU as demonstrated in the much smaller

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 15

coefficients with significant coefficients ranging from 0.08 to 0.51. The lag coefficient

appears to have the strongest influence in the second subperiod where ten of the fourteen

coefficients are in excess of 0.40.

The above results demonstrate how well the models fit the data and how well each

independent variable helps explain the movements of the return on each individual

currency. The thrust of this study, however, is to determine whether the volatility of the

currencies has changed with the increased development of the economic bloc. The

answers to these questions can be seen in Table 1-5. The second column notes the

average volatility for the individual currencies for the period 1979 through 1985. The

third column gives similar figures corresponding to the period from 1986 through 1992.

The fifth column reports the average daily volatility for each currency for the third and

final period, 1993 through April 1998.

(Insert Table 1-5 here)

The fourth column shows the percent change in volatility of the return on the

individual currencies with respect to the US dollar from the first period to the second

period. It is interesting to note that only four of the fourteen currencies experienced an

increase in their volatility. Of those four countries (Austria, Finland, Ireland and Italy),

Ireland’s percent change is very small (7.42 percent) and Italy’s is not much greater (14.2

percent). The remaining two countries experience important increases in volatility with

Austria’s increasing by 69 percent and Finland’s increasing by 9,114 percent. It is

interesting to note that these two currencies were not involved in the exchange rate

mechanism of the European Union at any time during this time period. Greece and

Portugal on the other hand, experience drastic decreases in the volatility of their

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 16

currency’s return, 98.99 percent and 99.95 percent, respectively. The rest of the countries

experience more moderate, but notable, decreases in their currencies’ volatility.

The sixth column reports the percent change in the individual currencies’

volatility from the second to the third time period. During this period only one country

experiences an increase in the volatility of its currency, Portugal. Of course, so much

volatility had been removed for Portugal from the first to the second period that even an

increase in volatility of 610 percent, as is the case here, still shows a large decline from

the first to the third period as noted in the seventh column. Aside from Portugal, all but

three countries, Ireland, Italy, and the UK, experience drops in the volatility of their

currency in excess of 50 percent. Finland’s and Sweden’s decrease the most with a 99.93

percent and 89.44 percent drop, respectively.

The seventh column is the most telling. It is interesting to see how the volatilities

have changed over the separate turning points in the level of integration of the European

Union, but what most people are looking for is the bottom line being what has changed

from then to now. All show some decrease with the exception of Ireland which shows

virtually no change at all (4.54 % increase in volatility). Some decreased quite

dramatically, with all but four currencies (Austria, Ireland, Italy, and UK) realizing a

volatility decrease in excess of 70 percent and 6 currencies (Denmark, Finland, France,

Greece, Portugal, and Sweden) realizing a decrease in volatility in excess of 80 percent.

These changes are displayed graphically in Figures 1-15 through 1-17.

(Insert Figures 1-15 through 1-17)

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 17

E. Conclusions

Here we have examined changes in the volatility of 14 European countries’

currency exchange rates as Europe has progressed toward a single economy. The

hypothesis we have is that as Europe experiences important events toward its

development as an integrated economic bloc the individual currencies of the affected

nations will become more stable. This stability will become manifest through decreased

volatility in exchange rates. The two events considered are (1) the declaration of a

program which became known as “Europe 1992” in 1986 and (2) the time at which this

program was scheduled to be completed in December 1992.

The findings of the empirical results of this section demonstrate that these

European currencies are generally well fitted by an AR(1) - EGARCH(1,1) model. Also

noted is that changes in the return in the individual currencies are very close to changes in

the return of the ECU and that this relation has apparently increased over time. Finally, it

is seen that for all but Ireland and Italy there has been a substantial decrease in currency

volatility as the time periods progress. This includes decreases in volatility ranging from

44 to 99 percent. This study has shown that the European Union may boast of at least

one more accomplishment. That accomplishment is that over the twenty years since the

introduction of the ECU, 12 of the 14 examined currencies have experienced notable

decreases in volatility.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 18

Chapter II

Volatility Changes in European Currency Exchange Rates

Due to EMS Announcements

A. Introduction

Many European nations have been committed for the last several years to

becoming a single market, not unlike the United States’ market. When stated this way it

is a very attractive idea. The United States has arguably the strongest market in the

world. The U.S. market is cohesive and is many separate countries’ largest trading

partner. Of course, emulating this is an attractive idea, however, many changes have

been made and many more need to be made for the European Union (EU) to reach this

goal.

Currently, European Monetary Union (EMU) is one goal of the proponents of a

single market that is under debate. It is frequently asked whether people believe there

will ever be a single currency for all of the nations in the EU. After having removed

several barriers to trade such as tariffs and duties and enacting similar laws regarding

local content and taxes, the EU has come a long way towards their goal. However, it is

argued that a single currency will facilitate trade both within and outside of EU. This has

costs attached to it. Many nations believe that they will lose sovereignty when they no

longer have control over how much money they are allowed to print. As it happens, they

really do not have much control now given that they are required by agreement to keep

the exchange rates of their currency within a certain range in relation to other countries

whose currencies participate in the Exchange Rate Mechanism (ERM). For this reason,

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 19

they currently have very little discretionary power over how much money they print

given that increasing or decreasing the money supply will obviously affect their exchange

rate. Regardless, the debate goes on.

Currency exchange has obvious implications for business. Any international

finance text will mention within the first five pages that 75 percent of U.S. companies

that do business outside of the U.S. have 100 or fewer employees (e.g. Madura, 1997,

p.4). Many of these smaller companies are not going to have the savvy to understand the

intricacies of the many exchange rates of the smaller countries of Europe. While all of

these currencies will trade directly with the U.S. dollar, given that it is a popular vehicle

currency, they will have fewer problems than if it were a small company in a small

country trying to trade with a company in another small country. However, there is a

certain amount of understanding that is required to effectively do business with many of

the smaller countries’ companies. Without this understanding it is much easier for a

small U.S. company to conduct business with a company in a larger country with whose

currency they are more familiar, such as Germany or U.K. There are several problems

that stem from this.

One problem is that the small U.S. companies may not be receiving the best deal

on the goods or services they are purchasing. This will lower their competitiveness.

Also, the smaller countries will not receive the business they rightfully deserve if they are

offering quality products at competitive prices. For these reasons the matter of exchange

rates within the EU is of great importance to the value of the firm. Were the process

simplified by a single currency, this could arguably increase both the competitiveness of

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 20

these smaller companies and likewise increase the competitiveness of the smaller

European countries.

It is generally accepted (at least in the popular business press) that when there is

good news for this single currency, the Euro, this will necessarily be bad news for the

Deutschemark. To our knowledge this has not been tested. In this study we attempt to

determine whether this is actually how the markets behave. Conversely, if good news for

the Euro is bad news for the mark, then good news for the Euro should be good news for

the weaker currencies whose countries’ economies will be strengthened by a single

European currency. The examined countries’ currencies are the Portuguese escudo, the

Italian lira, the Greek drachma, and the Spanish peseta. These countries are chosen

because they are frequently referred to as those which are making EMU difficult to attain.

The Italian lira was once removed from ERM due to Italy’s inability to keep the lira’s

exchange rate from fluctuating outside of its band. The Spanish peseta had similar

trouble that caused its bands to be widened more than those of other countries

participating in ERM did. This study examines whether announcements obtained from

the Wall Street Journal regarding the possibility of a single currency or the development

of a central banking system for the EU affect the volatility of these several currencies. It

is expected that announcements carrying good news for the Euro or the central banking

system will increase volatility in the mark’s exchange rate (as seen in French, Schwert,

and Stambaugh (1987)) and decrease volatility in the other currencies’ exchange rates.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 21

B. Literature Review

From previous literature that examines exchange rate behavior of the member

countries in the Exchange Rate Mechanism (ERM) in the European Monetary System,

we find three particular areas of study that are relevant to what is examined in this study.

First of all, a great deal of research is devoted to determining whether the Deutschemark

(DM) has as much influence on the exchange rates of other countries participating in

ERM of the EMS as is popularly believed. Along with this research is the study of

Germany’s actions such as monetary policy which will directly affect the value of the

DM and, therefore, indirectly affect the value of the other currencies, specifically those

participating in ERM. What is seen is that this vein of the literature is varied and quite

often contradictory.

Wyplosz (1989) finds that member countries that have greater restrictions

regarding monetary policy than other member countries of a fixed exchange rate system,

particularly ERM, have greater influence within the system. Given that Germany has

some of the most restrictive rules it will exert the most pressure or influence which will

enable Germany to dominate in this exchange rate system. MacDonald and Taylor

(1991) find similar influence. Their results show that ERM countries’ exchange rates,

both nominal and real, move together more in the long run than do countries’ currencies

in a floating exchange rate system. Their results suggest that this has been done through

monetary policy which has increasingly been modeled after the German standard in EMS

countries. German interest rates are found to dominate the interest rates in EMS

countries (Karfakis and Moschos, 1990). However, Katsimbris and Miller (1991)

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 22

determine that Karfakis and Moschos (1990) results are a function of the fact that the

study is too narrow and does not include important outside factors such as U.S. interest

rates which the later study finds to hold great influence.

Conversely, von Hagen and Fratianni (1990) dispute all of these findings and,

find, rather, that Germany is a very strong player, but suggest that to say that Germany

dominates is a gross overstatement. They do show it to be the least dependent nation of

the member countries, but they also witness this independence diminish over time.

A second area o f study that is relevant to the current study is seen in the numerous

efforts to model the behavior of the movement of exchange rates in ERM. Meese and

Rose (1990) use Locally Weighted Regression to test for nonlinear effects in fixed

exchange rate systems. They find no significant non-linearities except a few for the

French franc/ German mark rate. Vlaar and Palm (1993) examine the time-series

properties of exchange rates of the country currencies participating in ERM. They find

that the adjustments to ERM are captured by a Moving Average (1) - GARCH (1, 1) -

jump model.

Ball and Roma (1993) also try to find a good model of the exchange rates for the

currencies in ERM. They find that as EMU progresses, the ‘best’ model changes.

Initially a Brownian Motion process fits the data adequately, but in the later stages of

EMU they find that a mean reversion model is more appropriate. This suggests a single

currency is becoming a more likely outcome because this mean reverting behavior is

believed to be derived from the convergence of inflation and interest rates. Floating

currencies do not show mean-reverting behavior.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 23

To investigate the effects of the realignments of ERM in EMS countries Cheung,

et. al. (1995) use reduced rank cointegration. Their results support cointegration of the

exchange rates and therefore support purchasing power parity (PPP) between the many

countries. Contrarily, Edison and Fisher (1991) find that the artificially fixed exchange

rates were not cointegrated with prices, PPP does not hold, and that the weaker

economies may actually suffer due to ERM. The difference could be due to increased

efficiency of a maturing system or an increased acceptance of the possibility of a single

currency. Many of the previously mentioned studies find that the results have improved

over time, which could guide the EU toward a single currency.

Most recently, exponential GARCH (EGARCH), as developed in Nelson (1991)

is seen to provide an adequate representation of the volatility found in EMS countries’

currencies exchange rates (Hu, Jiang, and Tsoukalas, 1997). Due to the arrangements

inherent in EMS, there may be asymmetry between countries’ reactions to volatility

shocks. EGARCH provides a model specification which allows separate effects of good

and bad news along with a structure to examine persistence of the volatility.

The third area of study that is of particular relevance to what is being examined in

the present study, has to do with whether economic variables are converging, what might

be affecting them, and in what manner are they affected. The inflation and interest rates

in countries participating in ERM of the EMS and the U.K. are examined in Koedijk and

Kool (1992). They find that the ERM and its few adjustments are not bringing the rates

of the separate countries together to any great extent. Similarly, convergence between

these and other important economic variables is limited as seen in Beer and Knight

(1997). Koedijk and Kool (1992) do note that the countries, which are quick to act on

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 24

these economic differentials, such as the U.K., maintain more stability than those

countries that are slower to respond.

In examining the idea that increased currency substitution has a destabilizing

effect, Canzoneri and Diba (1993) find that the opposite is the case. If currency

substitution is stabilizing this makes EMU more viable. The authors note that if the

uncertainty in the system does not come from monetary policy, the witnessed stability

may be coming from a system other than currency substitution. However, this system

may itself be becoming less stable. If this is the case, the stability will then also

disappear in this system. If this is correct, the announcements examined in the present

study should reduce volatility for the DM. This is not what should be expected given the

reasoning suggested earlier from French, Schwert, and Stambaugh (1987) that bad news

induces increased volatility not decreased volatility.

Von Hagen and Neumann (1994) look at the variability in the real exchange rates

and find it to be decreasing. Of course, this is good news for those who support EMU.

The results are not as promising for Denmark, U.K., and Italy. However, Denmark

chooses not to support EU as a whole, U.K. has until recently been completely against

EMU since it removed itself from ERM in 1990, and Italy has had trouble keeping its

exchange rate within the limits of ERM and was involuntarily removed from ERM.

These events explain these particular countries not producing results similar to the

countries that are more directly involved.

In the previous chapter, we examine the changes in volatility of 14 European

currencies. In that study we witness a marked decrease in the volatility of these

currencies exchange rates from the inception of the European Currency Unit (ECU),

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 25

through the changes in the structure of the European Union, to the present in all but the

Irish pound and the Italian lira. This indicates that the progression of European unity has

had a positive and stabilizing effect on the exchange rates of these currencies.

The present study differs from all of these previous works in that it examines the

volatility of the different exchange rates. This has not been seen in the literature prior to

this work, except in Canzoneri and Diba (1993). They, however, examine different types

of events. It is proposed here that by examining the measures of relative volatilities in

the different exchange rates and looking for any difference in these volatilities around the

time of possibly important announcements regarding EMU we can measure whether a

single currency is good or bad news for each particular currency or if the currencies have

measurable, consistent responses at all.

C. Data and Methodology

C-i. Data

The data used in this study are the daily exchange rates of several European

currencies to the U.S. dollar. Specifically looked at in this paper are the German mark,

the Portuguese escudo, the Italian lira, the Greek drachma, and the Spanish peseta. The

reason these specific currencies are chosen from the many separate currencies in the

European Union is as follows: it is widely accepted conventional knowledge that any

good news for the Euro, the proposed name of the single currency in Europe, is bad news

for the German mark. The German mark is considered the strongest currency in the EU

and some evidence for and against this is seen in the previous literature. Furthermore, if

the Euro poses a threat to the stronger currencies in Europe, e. g., the mark, then it should

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 26

be considered good news for the weaker currencies. This study uses the southern

countries’ currencies since these countries are the ones most often suggested to bring the

most difficulty to the completion of the goals of the EU. Specifically, these countries

tend to not meet the guidelines set to enter into a single currency by the year 1999. Their

inflation, interest and unemployment rates are not meeting the standards, while many

northern countries are experiencing fewer of these difficulties as to the measures of

economic health. Daily exchange rate data is obtained from the U.S. Federal Reserve

Bank. Due to development of the ECU, the data begins January 3, 1979 and ends April

24, 1998. This leaves us with 4,850 observations for each currency with the exception of

the Greek drachma whose data begins April 13, 1981 and provides 4,279 observations.

The particular event dates to be examined in this study were obtained from an

investigation of the Wall Street Journal index. A search was undertaken to find all

announcements related to the single currency or a central banking system in EU. Once

located in the index, the articles were then obtained and examined to determine their

relevance and whether the news indicated is positive or negative in respect to the

actuality of a single currency or the development of a central bank, 47 articles were

found. O f course, many announcements were found to be unacceptable because they are

commentary in nature, 25 were removed. Remaining are 22 dates that are examined here

and presented in Table 2-1.

[Insert Table 2-1 here]

C-ii. Methodology

The method of examination used is to measure the average volatilities of the

month prior to the event and the month after the event and compare the percent change in

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 27

volatility. This method is taken from French, Schwert, and Stambaugh (1987) and

Schwert (1989). GARCH (1,1) models are used for the currencies to measure the daily

volatilities. Engle introduced the autoregressive conditional Heteroskedasticity (ARCH)

model in Engle (1982). This model allows the conditional variance to change over time

as a function of past errors. The strength of this model is that the conditional means and

variances can be estimated jointly using traditional specified models for economic

variables.

In this model, Yt is a random variable whose mean is given by Xtp (independent

variables) and is a linear combination of lagged endogenous and exogenous variables

included in the information set with p, a vector of unknown parameters.

Yt | o t.,~ N (X tp,ht)

ht = a0 + ZiaiEe2t.i (1)

£t= = Y ,- X tp

Bollerslev (1986) extends the ARCH process to GARCH (Generalized

Autoregressive Conditional Heteroskedastic), which allows for a more flexible lag

structure. Bollerslev points out that the extension of the ARCH process is very much like

the extension of the standard time series process to the general ARMA process.

The GARCH (p, q) regression model is obtained by

e t= Y ,- X tp

et.i | Ot-i N(0, h,) (2)

ht = a 0 + 2i=iqai62i + 2i=ipPiht-i

p > 0 q > 0

Where ao>0 aj > 0 i= l,...,q.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 28

Pi >0 i = 1,..., p.

For p = 0, the process reduces to the ARCH (q) process, and for p = q = 0 Sj is just white

noise. Bollerslev shows that the resulting GARCH (p, q) model is essentially a stationary

ARCH(q) process. We utilize the following GARCH model to study the impact of these

specific announcements on the exchange rate volatility.

Rt = Po + Pi Rt -l + P2 R-ECUt + st

s,., | 0),., N(0, ht) (3)

ht = ao + a ih n + a2S2t-i

Where Rt is defined as log (St/ St-i) * 100, where St is the spot exchange rate at time t (as

in Baillie and Bollerslev, 1989), R-ECU, as log ((ECU/US$)t/ (ECU/ US$),-0 * 100, and

ht is variance of st and is calculated recursively by a system of equations (3).

Bollerslev shows that in a GARCH (p, q) process the orders of p and q can be

identified by applying the traditional Box and Jenkins time series techniques to the

autocorrelations and partial autocorrelations for the squared process of et. Since the

autocorrelation and partial autocorrelation for the squared residuals from model (3) cut

off after lag one, we selected GARCH (1, 1) as the appropriate model. Bollerslev (1986)

also shows that GARCH (1,1) adequately fits many economic times series.

A limitation of the GARCH model described above is the conditional variance

responds to positive and negative residuals, st.i, in the same manner. However, empirical

evidence in financial time-series shows that there is a negative correlation between the

current returns and future return volatility. The GARCH model imposes the nonnegative

constraints on the parameters, cti and yi, while there are no restrictions on these

parameters in an exponential GARCH (EGARCH) model proposed by Nelson (1991). In

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 29

the EGARCH (1,1) model, the conditional variance, ht, is an asymmetric function of

lagged residuals 8t-i:

R t= Po + Pi Rt -i + P2 R-ECUt + s t (4)

ln(h,) = © + a i g (z,-i ) + yi ln(ht.i)

where g(zt) = 0zt + y[|zt| - E|zt|] and zt = e,/Vht. Consider the g(zt) function above. If z, is

positive then g(zt) is a linear function of the slope changes, zt, with slope (0 + y). If zt is

negative then the slope changes to (0 - y). Consequently, the conditional variance ht

responds asymmetrically to the sign of innovation zt.i.

D. Empirical Results

The estimates for the AR(1) - EGARCH(1, 1) model are given in Table 2-2,

Panel A. We observe that all relevant coefficients are highly significant and that the

amount of variation explained by the model is very high as seen in the R-square figures.

For Germany, a change in the return on the ECU is followed almost identically by the

German mark as observed by the coefficient equal to 1. This is interesting when one

notices that Germany has the highest coefficient for the return on the ECU and, therefore,

moves almost exactly as the ECU moves (given nothing else changes). Alternatively, the

remaining currencies have coefficients for the return on the ECU ranging from 0.82 for

Italy to 0.88 for Portugal. Thus, apparently, these currencies are not as strongly affected

by changes in the return on the ECU as is the German mark.

[Insert Table 2-2 here]

It also appears that the data is well fitted by the AR(1) - EGARCH(1, 1) model.

This can be observed both by the significant ai and yi coefficients in each of the five

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 30

models along with the high R-square levels. In addition to this, Panel B of Table 2-2

provides the diagnostics for each model. The Akaike Information Criteria (AIC) and the

Log Likelihood (LnL) are used to measure the appropriateness of the model for the given

data. Also, for non-linear time series models, the portmanteau Q-test statistics (Q) based

on standardized residuals (s t/Vht) are used to test for non-linear effects. The Q (10)

statistic cannot reject the null hypothesis of no nonlinear effects for up to lag 10 for any

of the 5 currencies. Thus it appears that the nonlinearity in the volatility series has been

successfully removed by our GARCH model specifications. Also reported is the

LaGrange multiplier test (LM) for ARCH disturbances proposed by Engle (1982) in

Panel B. The null hypothesis that the disturbances lack ARCH effects is not rejected.

The above results establish that the AR(1) - EGARCH(1, 1) model adequately fits

and measures the changes in the exchange rates of these five European currencies. We

now would like to examine the observed daily volatilites to determine if a relationship to

each of the above mentioned events and changes in the examined currencies’ volatilities

such as that suggested by French, Schwert, and Stambaugh (1987) exists. The results of

these tests are presented in Table 2-3. The average daily volatility for each currency is

examined for 20 days prior to each event (with the day prior to the event excluded given

that the announcement would be made the day prior to appearing in the Wall Street

Journal) and 20 days after each event (including the day prior to the event for the same

reason given above) are presented here. A twenty-day measure is used since each trading

month is approximately 20 days after considering holidays. In addition to this, the

percent change in average volatility from the time prior to the event to the time including

and subsequent to the event are calculated and presented here. Figures 2-1 through 2-5

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 31

display the daily volatility of the separate currencies over the time period examined.

Figures 2-6 through 2-10 display the change in volatility experienced by the separate

currencies around the event dates. These are provided so that one might more easily

observe the changes that occur around these dates.

[Insert Table 2-3 here]

[Insert Figures 2-1 through 2-5 here]

As can be seen in Table 2-3 and Figure 2-6, the German mark’s volatility shows

negligible change (change of less than 10 percent) in 2 of the 22 events, events 11 and 16,

is decreased in 10 events and is increased in 10 events. The increases are seen to be

greater in magnitude than are the decreases in that the average increase is 61.02 percent

and the average decrease is only 34.76 percent. Regardless of these observations, it is

difficult to claim that there is any recognizable pattern of volatility change for the

German mark, except that negative reactions appear stronger.

[Insert Figure 2-6 here]

Figure 2-7 and Table 2-3 display the changes in volatility for the Portuguese

escudo. For this currency we observe that of the 22 events 12 display decreases in the

volatility of the escudo. It should be noted that these decreases are on average of similar

magnitude to the increases in volatility. The average increase in volatility after the two

negligible changes of event 1 and 18 are excluded is 32.26 percent and the average

decrease in volatility is 39.18 percent. The number of changes in the opposite directions

is not proportional. The number of decreases is 50 percent greater than the number of

increases with 12 decreases and only 8 increases. This would appear to indicate that

news of the Euro is generally good news for the escudo.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 32

[Insert Figure 2-7 here]

Figure 2-8 and Table 2-3 display the results for the Italian lira. It is observed here

that 9 of the 22 event dates in effect show no effect given that the percentage change in

the lira’s volatility is less than 10 percent in either direction. Of the changes that are

greater than 10 percent, 9 are decreases in volatility and 4 are increases. The 4 observed

increases are generally substantially greater than the decreases as easily witnessed in

Figure 2-8. The average increase is 122.74 percent and the average decrease is only

38.09 percent. However, again there is no easily discemable pattern and the large

number of small changes would leave us to conclude that the lira is generally not strongly

affected by these announcements.

[Insert Figure 2-8 here]

Figure 2-9 and Table 2-3 display the percent changes in volatility of the Greek

drachma. For the drachma, only 3 of the event dates display a change of less than 10

percent in either direction, those are events 19, 21, and 22. Eight events display a notable

increase in volatility and 11 events display a decrease in volatility. If the one anomalous

change of 3,250 percent in event 7 and the 3 negligible changes are excluded, the average

changes both up and down are similar with increases averaging 36.93 percent and

decreases averaging 42.01 percent. Thus one might say that there are more decreases

than increases, but the average change in either direction is quite similar.

[Insert Figure 2-9 here]

Figure 2-10 and Table 2-3 offer the results for the Spanish peseta. We observe

that events 5, 12, and 22 show negligible effect given that the percent change is less than

10 percent in either direction. Of the remaining events, 9 display a decrease in volatility

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 33

and 10 display an increase in volatility. Although there are a few extreme increases in

volatility around an event date, when the one extreme increase of 197 percent and the

negligible changes are excluded the average change in either direction is similar with

increases averaging 47.83 percent and decreases 43.02 percent. Two of the three event

dates around a negative announcement, events 8 and 13, provide large increases in

volatility of 28.16 percent and 81.22 percent. However, these are not isolated incidents

of increase. Each announcement date is as likely to provide an increase in volatility as a

decrease and the magnitude is generally not very different, thus it again appears that no

discemable pattern may be found in the changes in volatility of this currency around

these particular event dates.

[Insert Figure 2-10 here]

One more interesting observation from Table 2-3 is that several of the events

elicit similar reaction across countries, rather than a different reaction from the weaker

countries than Germany. It is interesting to note that reactions were similar for 8 of the

first 11 events across countries in that all currencies’ volatilities changed in the same

direction, but only 2 of the 11 later events elicit similar reactions across countries.

Events 1,2,3, 6, 7, 9,10, and 11 all show changes in the same direction across countries

in the first 11 events. Only events 15 and 21 elicit similar reactions across the countries

for the latter 11 events.

E. Conclusions

This paper has examined five separate European currencies, the German mark, the

Portuguese escudo, the Italian lira, the Greek drachma, and the Spanish peseta, to

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 34 / determine if there is any noticeable change in the volatility of these currencies’ exchange

rates after an announcement pertaining to a single currency in Europe. Initially, we find

that a AR(1) - EGARCH(1,1) model is well fitted to the data.

As to the effects noticed after the announcements, Germany and Spain experience

a similar amount of increases as decreases. Germany’s increases in volatility appear to

be much more severe than the decreases. This could imply something that has been

supposed before that negative news is more strongly reacted to than positive. Italy also

displays much stronger reactions to negative news as implied by a much stronger

increases in volatility than the more frequent decreases.

If it is the case that bad news elicits a greater reaction than good news and bad

news for Portugal, Greece and Spain’s results would imply that whichever events are

perceived as bad news this news is not as bad as the good news is good. While the model

fits the data well and does a more than adequate job of explaining the variation in returns,

we are not able to readily explain what reaction any particular will have to the EMU

announcements. This could be due to the fact that the fine details of the effects of each

announcement’s content are either missing from theWall Street Journal’s article or are

not completely understood by the researcher.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 35

Chapter III

Volatility Changes in European American Depository Receipt Returns: Evidence from

the NASDAQ Market

A. Introduction

The question of whether exchange rates affect stock prices, or vice versa, is an old

one. The premise is a sensible one. First, one must consider what is assumed to

constitute the value of a stock. A stock’s value is the present value of its future cash

flows. The value of these future cash flows will obviously be affected by exchange rates

given that exchange rates will be a determinant of the real value of the nominal amount of

those future cash flows. What interests us in this study is whether European American

Depository Receipts (ADRs) are affected by announcements concerning a single currency

in Europe’s likelihood, composition, and timing. Although ADRs have been seen to not

behave exactly the way stocks do, they are very similar in concept.

As the European Union strives to develop a single currency for the several

nations, all aspects of the economies of the nations will be affected. In order to become a

single market, the separate European nations have accepted many changes in the manner

business is conducted between the member nations. Barriers to trade have been lessened

or removed to a great extent. Issues are debated and resolved over some of the smallest

details. One issue which remains in debate is the idea of a single currency. It has been

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 36

decided that this process will take place in what has come to be known as a “Europe of

two speeds.” That is 11 countries have been decided to originally meet the criteria to join

a single currency will do so in 1999. Others will be put on the waiting list to be allowed

to join sometime shortly after as their relevant economic criteria become closer to those

required for membership.

These occurrences undoubtedly have some impact on the value of the companies

of the separate countries. Those countries, which are not allowed to enter into the single

currency, will continue to participate in the Exchange Rate Mechanism (ERM). In doing

so they will continue to keep their exchange rates in line with what is expected and

balance out whatever other economic situations they have that are keeping them out of

the single currency.

European Monetary Union is currently a controversial topic. Economists,

business people, and politicians alike argue over whether it should happen, whether it can

happen, and whether it will happen. There are many arguments on either side. Over

twenty years ago, many leaders of the nations of Europe developed the goal of molding

all of Europe into a single market. Much progress has been made toward this goal

including the lowering of trade barriers such as tariffs and duties between the member

nations. This has enabled goods and services to cross country boundaries with much

greater ease. Also, much progress has been made in unifying Europe in terms of

economic measurements. Similar monetary and fiscal policies, both in relation to the

Exchange Rate Mechanism (ERM) and more simply in relation to achieving similar

inflation, interest, and unemployment rates, between the member nations are being

applied. However, there is still the question of monetary union.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Will monetary union occur for the excluded nations? That is to be seen. The

concern of the present study is not to determine whether monetary union will occur or

even should it occur, but rather to measure if the opportunity or threat of monetary union

elicits a reaction from investors in equity holdings of European companies. In this paper,

we examine whether investors in European companies, by way of American Depository

Receipts, display a noticeable reaction to announcements of progress toward both a single

currency in the member nations of the European Union and the development of a central

banking system for this single currency.

The paper is laid out as follows: in the next section, previous literature related to

this subject is reviewed. In section C the data and methodology are discussed. The fourth

section presents the results and discussion. The fifth and final section offers conclusions

of the findings.

B. Literature Review

Much research has been done in the area of the relationship of changes in foreign

exchange rates and stock prices or ADRs. The results, however, have been somewhat

mixed. Thomas (1988) finds that 10 of 15 countries examined show a positive

correlation between equity prices and the dollar value of the local currency. However,

these correlations are low and generally not significant. Ma and Kao (1990) examine

both exchange rate changes and exchange rate levels in relation to equity prices. They

find that exchange rate levels’ relationship to stock market indexes is positive and that

exchange rate changes are negatively related to the stock market indexes. The exchange

rate levels, however, are seen to have a greater influence on stock indexes.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 38

In an important paper with regard to Arbitrage Pricing Theory, Roll (1992)

studies the volatility of stock market indices and finds that one factor of significant

influence is exchange rates. These indexes are not as strongly influenced by exchange

rates as they are by the country’s industrial structure, but the influence is still strong and

worth noting.

Ajayi and Mougoue (1996) study the long- and short-term relationship of stock

indexes to the exchange rate of the country. They find that the two series are co

integrated although long- and short-term properties differ. Najand and Yung (1997),

using futures contracts, find a significant negative effect of stock index futures on foreign

exchange futures which implies that a strong stock market could make for a strong

currency. Given all of these findings it is clear that exchange rates and stock prices are

related to one another. This study, however, is unique. Here we elect to examine

whether announcements found in the Wall Street Journal affect equity prices of

companies from European Union member countries.

C. Data and Methodology

C-i. Data

The data used in this study are the daily prices of American Depository Receipts

(ADRs) of companies located in countries which are members of the European Union.

This data was collected from NASDAQ. ADRs are chosen for two reasons: first, they are

unique in nature in that they are not stock themselves, but rather a certificate of

ownership issued by U.S. banks which represent a claim to underlying foreign securities

(usually common stock of the company in question). Secondly, Wahab and Khandwala

(1993) determine that ADRs dominate simple foreign stocks in that they provide similar

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 39

returns to that of foreign stock, but offer more diversification of risk. This gives evidence

that ADRs behave, at least to some small degree, differently on the market.

In an attempt to simultaneously maximize both the number of companies

examined and the length of examination, it is decided that only ADRs that have traded on

NASDAQ for at least four years will be used. Thirty-two such companies are found.

Due to the choice of companies and availability of data, daily prices are collected from

September 1, 1993 through September 26, 1997. This provides the study with 1029

observations.

The particular event dates to be examined in this study were discovered by an

investigation of theWall Street Journal index. A search was undertaken to find all

announcements related to the single currency or a central banking system in EU. Once

located in the index, the articles were then obtained and examined to determine their

relevance and as to whether the news indicated is positive or negative with respect to the

actuality of a single currency or the development of a central bank, 47 articles were

found. Of course, many announcements are found to be unacceptable because they are

commentary in nature, 25 were removed. The length of time the ADR prices are

available also disqualified many of the remaining announcement dates, all

announcements prior to September 1,1993 (12) were removed. Remaining are 10 dates

that are examined here and presented in Table 3-1.

[Insert Table 3-1 Here]

C-ii, Methodology

The method of examination employed is that we calculate the average of the daily

volatility for each country for twenty days prior to the event and twenty days after the

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 40

event and compare the percent change in volatility. Of course, the day immediately prior

to the date of the announcement is included in the twenty days after since the

announcement will appear in The Journal the day after the news breaks. This method of

measurement is taken from French, Schwert, and Stambaugh (1987) and Schwert (1989).

A GARCH (1, 1) model is used for each country after that country’s respective ADRs are

combined into an equally weighted portfolio to measure the daily volatilities of the

country portfolio’s returns. Some countries have several ADRs that fit our criteria and

were therefore obtained for this study and other countries only have one or two ADRs

that fit our criteria. Finland, France, Greece, and Luxembourg each have only one ADR.

The Netherlands has two ADRs. Ireland has four ADRs. Sweden has five ADRs. The

U.K. has seventeen ADRs.

Engle introduced the autoregressive conditional Heteroskedasticity (ARCH)

model in Engle (1982). This model allows the conditional variance to change over time

as a function of past errors. The strength of this model is that the conditional means and

variances can be estimated jointly using traditional specified models for economic

variables.

In this model, Yt is a random variable whose mean is given by Xtp (independent

variables) and is a linear combination of lagged endogenous and exogenous variables

included in the information set Ot.i with p, a vector of unknown parameters.

Yt |o t.,~N(Xtp,ht)

ht = ao + Ei

et= =Yt-X,p

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 41

Bollerslev (1986) extends the ARCH process to GARCH (Generalized

Autoregressive Conditional Heteroskedastic), which allows for a more flexible lag

structure. Bollerslev points out that the extension of the ARCH process is very much like

the extension of the standard time series process to the general ARMA process.

The GARCH (p, q) regression model is obtained by

e, = Yt -X ,p

s,., I O,., N(0, ht) (2)

ht = a 0 + £i=iqociS2t-i + 2i=ipPiht.i

p > 0 q > 0

Where ao>0 a\ > 0 i = l,...,q.

Pi > 0 i=l,...,p.

For p = 0, the process reduces to the ARCH (q) process, and for p = q = 0 Sj is just white

noise. Bollerslev shows that the resulting GARCH (p, q) model is essentially a stationary

ARCH(q) process. We utilize the following GARCH model to study the impact of these

specific announcements on the ADR price volatility:

Rit= Po + PiRit-i + p2R-NASDAQ+ G[

St., |

ht = ao + aiht-i + 0C2S2t-i

Where Rjt is the log of the current country portfolio value divided by the lag of the

countiy portfolio value times 100 for each country under examination (i.e., the log return

of the portfolio of ADRs in a country), R-NASDAQt is defined as log (NASDAQt/

NASDAQt-i) * 100 (i.e., the log return of the NASDAQ index), and ht is the variance of

st and is calculated recursively by a system of equations (3).

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 42

Bollerslev shows that in a GARCH (p, q) process the orders of p and q can be

identified by applying the traditional Box and Jenkins time series techniques to the

autocorrelations and partial autocorrelations for the squared process of et. Since the

autocorrelation and partial autocorrelation for the squared residuals from model (3) cut

off after lag one, we selected GARCH (1, 1) as the appropriate model. Bollerslev (1986)

also shows that GARCH(1,1) adequately fits many economic times series.

A limitation of the GARCH model described above is that the conditional

variance responds to positive and negative residuals, sn, in the same manner. However,

empirical evidence in financial time-series shows that there is a negative correlation

between the current returns and future returns volatility. The GARCH model imposes the

nonnegative constraints on the parameters,a\ and yi, while there are no restrictions on

these parameters in an exponential GARCH (EGARCH) model proposed by Nelson

(1991). In the EGARCH (1, 1) model, the conditional variance, ht, is an asymmetric

function of lagged residuals 8t.i:

Rt = Po + Pi Rt-i + P2 R-ECUt + Et (4)

ln(ht ) = co + cxi g (zt.i ) + yi ln(hn)

where g(zt) = 0zt + y[|zt| - E|zt|] and zt = et/Vht. Consider the g(zt) function above. IfZt is

positive then g(zt) is a linear function of the slope changes, zt, with slope (0 + y). If zt is

negative then the slope changes to (0 - y). Consequently, the conditional variance, ht,

responds asymmetrically to the sign of innovation zt.i.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 43

D. Empirical Results

The GARCH estimates for each country are displayed in Table 3-2. In Panel A of

this table one can observe that the return on NASDAQ (R-NASDAQ) is a significant

indicator for all countries, except Ireland. Interestingly, Finland has a negative

relationship to NASDAQ while all others have a positive relationship. Of those countries

with a positive relationship the coefficients vary from 0.126 for the Netherlands to 0.483

for Sweden. Four of these six countries have a coefficient of 0.31 or greater. This

indicates that changes in the NASDAQ index are moderately reflected in concurrent

changes in the country portfolios of ADRs. It should also be noted that the lag for each

portfolio is highly significant for all countries except France and Luxembourg. The

coefficients again vary to a large extent with the Netherlands and Sweden having

significant negative coefficients of approximately -0.5 for each and the other four

significant coefficients ranging from 0.03 for Finland to 0.21 for Greece. This indicates

that the lag is only a mild indicator of the current return on each portfolio.

(Insert Table 3-2)

Table 3-2 also displays that the AR(1) - EGARCH(1, 1) model fits the data well.

This can be observed in the highly significant cti and yi coefficients. Panel B of Table 3-

2 offers the diagnostics. Here the Akaike Information Criteria (AIC) and the Log

Likelihood (LnL) are used to measure the appropriateness of the model for the given

data. Also, for non-linear time series models, the portmanteau Q-test statistics (Q) based

on standardized residuals (et/Vht) are used to test for non-linear effects. The Q (10)

statistic cannot reject the null hypothesis of no nonlinear effects for up to lag 10 for any

of the 8 countries’ portfolio returns. Thus it appears that the nonlinearity in the volatility

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 44

series has been successfully removed by our GARCH model specifications. Also

reported in Panel B is the LaGrange multiplier test (LM) for ARCH disturbances

proposed by Engle (1982). The null hypothesis that the disturbances lack ARCH effects

is not rejected.

Table 3-3 shows the percent change in average daily volatility for each country’s

portfolio of ADRs for each announcement. The change in volatility is measured as the

percent change of the average daily volatility from the twenty days prior to the

announcement to the twenty days after the announcement. Twenty days are chosen since

on average a month includes approximately twenty trading days when holidays are

considered. Of course, the day prior to the announcement is included in the post

announcement average since the announcement will appear in theWall Street Journal the

day after the news breaks.

(Insert Table 3-3 here)

Finland, France, and the Netherlands show no notable change in volatility around

any of the announcement dates. For Finland and France this could be understandable in

that they each have only one ADR in their portfolio which may not be affected by such

events. However, this assumption brings up the question as to why Greece and

Luxembourg do show notable change in the volatility of their portfolios yet only have

one ADR in their respective portfolios. Figures 3-9 through 3-16 give a graphical

depiction of the percentage changes in volatility for each country portfolio.

(Insert Figures 3-9 through 3-16 here)

Of the notable changes for Greece, events 1, 2, 5, and 7 show a decrease in

volatility or a positive response and events3,4, 6, 8, and 9 display an increase in

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 45

volatility or a negative response. Greece’s greatest volatility changes occurred for events

7 (decrease) and 8 (increase). Event 7 describes an increase in faith in the Euro by the

Swiss. Event 8 describes the French and German governments increasing their

commitment to a single currency. It appears that either these changes are unrelated to the

announcements considered or there are changes in the attitude toward a single currency in

Europe in Greece. This could be due to Greece having difficulty maintaining compliance

requirements for participation in ERM.

Ireland has two events, 3 and 5, that show no notable response. Of the remaining

eight events 1, 7, 8 and 10 display decreases in volatility and events 2,4, 6, and 9 display

increases in volatility. Ireland experiences the greatest changes in volatility during events

10 (decrease) and 6 (increase). Event 6 describes Germany uncharacteristically issuing

short-term debt denominated in ECU. Event 10 describes how the German chancellor,

Helmut Kohl, insists on revaluing gold reserves in favor of European Monetary Union.

Like Greece, Ireland’s portfolio seems to be affected somewhat randomly by these

announcements.

Luxembourg’s portfolio contains a single ADR, but still shows many changes in

volatility. The only announcement for which there was no notable change in volatility

for Luxembourg is event 3. For the notable changes, 6 o f the 9, events 2, 5, 6, 7, 8, and 9,

are decreases in volatility and events 1 and 10 display increases in volatility. This would

make it appear that a single currency in Europe is viewed mostly positively in

Luxembourg. This stands to reason since Luxembourg has voluntarily pegged its

currency with Belgium and the Netherlands for some time. Luxembourg experienced the

greatest changes in volatility around events 6 (decrease) and 1 (increase). Event 1, a

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 46

negative announcement, describes the beginning of an ERM currency crisis which forces

Spain out of ERM. Event 6 shows the resolve on Germany’s part to establish a single

currency by denominating short-term debt in the ECU.

Sweden has three event dates where virtually no change is witnessed, events 2, 6,

and 9. Of the other seven, four are decreases in volatility, events 1, 4, 5, and 7 and three

are increases in volatility, events 3, 8, and 9. Again, this country’s portfolio appears to

react around the time of these announcements, but the outcome is unpredictable. The

changes in volatility for Sweden are also somewhat small ranging from -26.2 percent for

event 7 to 29.6 percent for event 3. Event 7 describes Switzerland increasing their

support for the ECU in order to decrease the strengthening of their own currency. Event

3 establishes a process by which the European Union will implement a single currency.

Only five of the announcements had a notable effect on the British portfolio,

events 1,2, 5, 6, and 7 indicate very little change in volatility. As for the remaining five

events, two witness decreases in volatility, events 4 and 5, and three are affected

negatively, events 3, 8, and 9. The greatest changes around any of these events are

observed around events 4 (decrease) and 8 (increase). Event 4 announces that the 15

members agree upon a new name for the single currency. Event 8 explains a display of

increased support of the single currency by the French and German governments. This

portfolio contains 17 ADRs and is therefore the largest portfolio. This could be a well

developed portfolio that could weather the storm and not be as affected by these

announcements given that some companies would find a single currency good news and

others would not.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 47

E. Conclusions

In this study we have examined several different European countries’ American

Depository Receipts (ADRs) to determine whether announcements of developments

toward a single currency in Europe have an effect on them. The ADRs are combined into

equally weighted portfolios by country of origin. We see that while many of these

country portfolios witness percent changes in volatility greater than 10 percent around

each event date, there is no obvious pattern for the combination of countries.

O f the 8 countries examined, Luxembourg has the most notable results. Of the

nine events for which there is a notable change in volatility, six are decreases. Also,

Luxembourg’s reactions are among the greatest in percentage changes ranging from -

43.4 percent to 46.3 percent and of these 6 are changes of 20 percent or more in either

direction. Although our results are somewhat inconclusive, it is still interesting to note

which countries’ ADRs are affected and which are not. Another interesting note is that

Greece, Ireland and the U.K. all reacted negatively (an increase in volatility) around

event 9. Event 9 announces the discussion of putting off a single currency for another

year to allow more time for nations to comply to requirements for entry into the single

currency. This could lead to further research in the area of ADRs, which has been less

researched than other similar areas.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 48

References

Abuaf, N. and P. Jorion (1990), “Purchasing power parity in the long run,”Journal o f Finance, 65, 157-174.

Ajayi, R. A. and M. Mougue (1996), “On the dynamic relation between stock prices and exchange rates,”Journal o f Financial Research, 19, 193-207.

Argimon, I. and J.M. Roldan (1994), “Saving, investment, and international capital mobility in EC countries,” European Economic Review, 38, 59-67.

Baillie, R.T. and T. Bollerslev (1989), “The message in the daily exchange rates: A conditional-variance tale,”Journal o f Business and Economic Statistics, 7, 297- 305.

______(1990), “ A multivariate generalized ARCH approach to modeling risk premia in forward foreign exchange rate markets,”Journal o f International Money and Finance, 9, 309-324.

Ball, C. A. and A. Roma (1993), “A jump diffusion model for the European Monetary System,” Journal o f International Money and Finance, 12, 475-492.

Bollerslev, T. (1986), “Generalized autoregressive conditional heteroscedasticity,” Journal o f Econometrics, 31, 307-327.

______(1987), “A conditional heteroskedatic time series model for speculative prices and rates of return,” Review o f Economics and Statistics, 69, 542-547.

______(1990), “Modeling the coherence in short-run nominal exchange rates: A multivariate generalized ARCH model,” Review o f Economics and Statistics, 72, 498-505.

Bollerslev, T., R.Y. Chou, and K.F. Kroner (1992), “ARCH modeling in finance: A review of the theory and empirical evidence,”Journal o f Econometrics, 52, 5-59.

Canzoneri, M. B.and B.T. Diba (1993)., “Currency substitution and exchange rate volatility in the European Community,”Journal o f International Economics, 35, 351-365.

Cheung, Y. W., H.G. Fung, K.S. Lai, and W.C. Lo (1995),” Purchasing power parity under the European Monetary System,” Journal o f International Money and Finance, 14, 179-189.

Chinn, M.D. (1991), “Some linear and non-linear thoughts on exchange rates,”Journal o f International Money and Finance, 10, 214-230.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 49

Choi, J.J. (1986), “A model of firm valuation with exchange exposure,”Journal o f International Business Studies, 17, 153-160.

Diebold, F.X. (1988), Empirical Modeling o f Exchange Rate Dynamics, Springer Verlag: New York, NY.

Edison, H. J.and E. Fisher (1991), “A long-run view of the European Monetary System,” Journal o f International Money and Finance, 10, 53-70.

Engle, R.F. (1982), “Autoregressive conditional heteroscedasticity with estimates of the variance of U.K. inflation,”Econometrica, 50, 987-1008.

Engle, R.F. and T. Bollerslev (1986), “Modeling the persistence of conditional variances,” Econometric Review, 5, 1-50, 81-87.

Engle, R.F., T. Ito, W-L. Lin (1990), “Meteor showers or heat waves? Heterskedastic intra daily volatility in the foreign exchange market,”Econometrica, 58, 525-542.

Engle, R.F., D.M. Lilien, R.P. Robins (1987), “Estimating time varying risk premia in the term structure: The ARCH-M model,” Econometrica, 55, 391-407.

French, G., G.W. Schwert,and R.F. Stambaugh (1987), “Expected stock returns and volatility,”Journal o f Financial Economics, 17, 3-29.

Friedman, D. and S. Vandersteel (1982), “Short-run fluctuations in foreign exchange rates,” Journal o f International Economics, 13, 171-186.

Froot, K. and M. Obstfeld (1991), “Exchange rate dynamics under stochastic regime shifts: A unified approach,” Journal o f International Economics, 31, 203-229.

Helg, R., P. Manasse, T. Monacelli, and R. Rovelli (1995), “How much (a)symmetry in Europe? Evidence from industrial sector,”European Economic Review, 39, 1017- 1041.

Hseih, D. (1988), “The statistical properties of daily foreign exchange rates: 1974-1983,” Journal o f International Economics, 24, 129-145.

Hu, M.Y., C.X. Jiang, C. Tsoukalas (1997), “The European exchange rates before and after the establishment of the European Monetary System,”Journal o f International Financial Markets, Institutions, and Money, 235-253. 1,

Jorion, P. (1990), “The exchange-rate exposure of U.S. multinationals,”Journal o f Business, 63, 331-345.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 50

Karfakis, J. C. and D. M. Moschos (1990), “Interest rate linkages within the European Monetary System: A time series analysis,” Journal o f Money, Credit and Banking, 22, 388-394.

Katsimbris, G. M. and S. M. Miller (1993), “Interest rate linkages within the European Monetary System: Further analysis,” Journal o f Money, Credit and Banking, 25, 771-779.

Koedijk, K. G. and C. J. M. Kool (1992), “ Dominant interest and inflation differentials within the EMS,” European Economic Review, 36, 925-943.

Krugman, P. (1991), “Target zones and exchange rate dynamics,” Quarterly Journal o f Economics, 106, 669-682.

Ma, C. K. and G.W. Kao (1990), “On exchange rate changes and stock price reactions,” Journal o f Business and Finance & Accounting, 17, 441-449.

MacDonald, R. and M.P. Taylor (1991), “Exchange rates, policy convergence, and the European Monetary System,”Review o f Economics and Statistics, 73, 553-559.

______(1994), “The monetary model of the exchange rate: Long-run relationships, short-run dynamics, and how to beat a random walk,” Journal o f International Money and Finance, 13, 276-290.

Madura, J. (1997), International Financial Management, Fifth Edition, South Western, Cincinnati, Ohio.

Meese, R.A. and K. Rogoff (1983), “ Empirical exchange rate models of the Seventies: Do they fit out of sample?”Journal of International Economics, 14, 13-24.

Meese, R.A. and A.K. Rose (1990),” Nonlinear, nonparametric, nonessential exchange rate estimation”, American Economic Review, 80, 192-196.

Mundell, R. (1961), “A theory of optimum currency areas,” American Economic Review, 52, 657-665.

Mussa, M. (1979), “Empirical regularities in the behavior of exchange rates and theories of the foreign exchange market,”Carnegie-Rochester Conference Series on Public Policy, 11, 9-57.

Najand, M. and K. Yung (1997), “Price dynamics among exchange rates, stock index, and Treasury Bonds in futures markets,” Advances in Investment Analysis and Portfolio Management, 4, 65-76.

Nelson, D.B. (1997), “Conditional heteroskedasticity in asset returns: A new approach,” Econometrica, 2,347-370.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 51

Phylaktis, K. and Y. Katsimatis (1994), “Does the real rate follow a random walk? The Pacific Basin perspective,”Journal o f International Money and Finance, 13,476- 495.

Poole, W. (1970), “Optimal choice of monetary policy instruments in a simple stochastic macro model,” Quarterly Journal o f Economics, 84, 197-216.

Pozo, S. (1987), “The ECU as international money,”Journal of International Money and Finance, 6, 195-206.

Roll, Richard (1992), “Industrial structure and the comparative behavior of international stock indices,” Journal o f Finance, 47, 3-41.

Rose, A.K. and L.E.O. Svensson (1994), “European exchange rate credibility before the fall,” European Economic Review, 38, 1185-1216.

Schinasi, G.J. and P.A.V.B. Swamy (1989), “The out-of-sample forecasting performance of exchange rate models when coefficients are allowed to change,”Journal o f International Money and Finance, 8, 375-390.

Schwert, G. W. (1989), “Why does the stock market volatility change over time?,” Journal o f Finance, 44,1115-1153.

Sutherland, A. (1995), “Monetary and real shocks and the optimal target zone,” European Economic Review, 39, 161-172.

Taylor, S.J. (1990), “Modeling stochastic volatility,” Unpublished manuscript (Department of Accounting and Finance, University of Lancaster, Lancaster)

Thomas, L. R., Ill (1988), “Currency risks in international equity portfolios,”Financial Analysts Journal, 44, 68-71.

Vlaar, P.J.G. and F.C. Palm (1993), “The message in weekly exchange rates in the European Monetary System: Mean reversion, conditional heteroscedasticity, and jum ps,” Journal of Business and Economic Statistics, 11,351 -360.

von Hagen, J. and M. Fratianni (1990), “German dominance in the EMS: Evidence from interest rates,” Journal o f International Money and Finance, 9,, 358-375.

von Hagen, J. and M.J.M. Neumann (1994), “Real exchange rates within and between currency areas: How far away is EMU?,” Review o f Economics and Statistics, 76, 236-244.

Wahab, M. and A. Khandwala (1993), “Why not diversify internationally with ADRs?” Journal o f Portfolio Management, 19 (Winter 1993), 75-82.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 52

Wyplosz, C. (1989), “Asymmetry in the EMS: Intentional or systemic?”European Economic Review, 33, 310-32

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 1-1 o o o o o o o o p o

CO 10 CO cvi 1.00

Figure 1-3 Volatility: Danish kroner Figure 1-4 Volatility: Finnish markka 0.00 50000.00 100000.00 150000.00 200000.00 300000.00 250000.00 350000.00 400000.00

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 1-6 > _re o lity: German mark

/ 14/96 4 Figure 1-7 Volatility: Greek drachma 0.00 100000000.00 300000000.00 200000000.00 500000000.00 400000000.00 700000000.00 600000000.00 800000000.00

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 20.00

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 1-11 Volatility: Portuguese escudo 0.00 200000.00 400000.00 800000.00 600000.00 1200000.00 1000000.00 1400000.00

Figure 1-12 Volatility: Spanish peseta

20.00 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 1-13 Volatility: Swedish kroner

1000.00 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 1-14 > g CO JZ TJ o II) Q. o 3 c VO vo

67 Sweden Figure 1-15 74 PercentChange in Exchange Rate Volatility from 1979-1985 to 1986-1992 % % % % 00 00 00 00 . . . . *The Finnish markaa's volatility increased by 9115%. 0 50.00% 75.00% 25.00% 50.00% -25.00% -50.00% -75.00% 100 125.00% 100 175.00% 200 -

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

68 Swedei Figure 1-16 PercentChange in Exchange RateVolatility from 1986-1992 to 1993-1998 % % % % 00 00 . 00 . 00 . . 0 *The Portuguese escu d o 's volatility increased by 611% 50.00% 25.00% 75.00% 50.00% -75.00% -25.00% 100 175.00% 150.00% 125.00% 100 - 200

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

69 Sweden Figure 1-17 Percent Change in Exchange RateVolatility from 1979-1985 to 1993-1998 % % % % % 00 00 00 00 . 00 . . . . 0 80.00% 20 20 80.00% 40.00% 60.00% -60.00% -40.00% 100 100 -

Figure 2-1 > o ro lity: German mark 98/ 98/ i/I 08/i/1 4/14/98 71 Figure 2-2 Volatility: Greek drachma 0.00 100000000.00 300000000.00 200000000.00 400000000.00 500000000.00 700000000.00 600000000.00 800000000.00

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2-4 Volatility: Portuguese escudo f-> f-> to 0.00 200000.00 400000.00 600000.00 800000.00 1000000.00 1200000.00 1400000.00

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2-5 Volatility: Spanish peseta 0.00 20.00 60.00 40.00 80.00 100.00 120.00 140.00 160.00

Figure 2-6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 2-7 erdcdwt priso o h cprgtonr Frhrrpouto poiie ihu permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 2-8 Percent Change per Event: Italy 78 C 'J O-J CO CD Figure 2-9 PercentChange per Event: Greece cc r-- co cr> *The change in the volatility ofthe G reek drachm a for event 7 is 3250%. % % % % % % % 00 00 00 00 00 00 00 ...... 0 20 20 80.00% 60.00% 40.00% -40.00% -60.00% -80.00% - 100 100 140.00% 120 180.00% 160.00% 200 -

Figure 2-10 Percent Change per Event: Spain Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 3-1 > o a lity: Finnish ADRs / / 96/2/!! 6 / / /112 CO © Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 3-2 itv: French ADRs Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 3-3 > O o < 0£ B o £ a> w 63i1 96/3/i /.6A;/ 96 / 3/11 ii 40.00

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 11/2/97 84 Figure 3-5 Volatility: Luxembourg ADRs 14.00 16.00

Figure 3-6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 3-7 > (0 £ < a 0£ o « I a> M

0.50 % v ■ % % % % \ % % >%> % \ % % % % % Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 3-8 Volatility: British ADRs Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 3-9 Percent Change per Event: Finland Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 3-10 Percent Change per Event: France Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 3-11 Percent Change per Event: Greece Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 3-12 Percent Change per Event: Ireland Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 3-13 Percent Change per Event: Luxembourg Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 3-14 Percent Change per Event: Netherlands Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 3-15 Percent Change per Event: Sweden Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without prohibited reproduction Further owner. copyright the of permission with Reproduced

Figure 3-16 Percent Change per Event: UK 96 R* 0.5819 0.4501 0.6690 0.6047 e (0.0344) (0.0308) (0.0449) 0.0911250.126259 0.4206 0.4196 0.648623 0.5153 (0.0389)*** (0.0271)*** (0.0394)*** (0.0266)*** (0.0427)*** (0.0433)*** (0.0365)*** (0.0306)*** (0.0340)*** -0.027477 0.6040 -0.231909 0.6210 Yi 0.664493 0.983600 -0.208680 0.6265 0.969404 0.362646 0.5879 0.870499 0.081354 0.4278 0.977861 0.958258 0.236416 0.5237 0.993889 0.037145 0.707698 0.391705 0.878788 -0.012163 0.967218 -0.147264 (0.0209)*** (0.0375)*** (0.0122)*** (0.00271)*** (0.00295)*** a, 0.344473 (0.0347)*** (0.0189)*** (0.0210)*** c0 (0.0356)*** (0.00583)** (0.0114)*** (0.0415)*** (0.00721)*** (0.0233)*** (0.0358)*** -0.021141 0.321221 -0.052304 0.330296 (0.00554)*** (0.0183)*** Table 1-1 Table Panel A 1-1 R,1 0.360319 0.399419 0.007760 0.197554 0.293971 0.347336 -0.410133 0.560868 0.380595 -0.175579 0.556434 (0.0171)*** (0.0481)*** (0.0310)*** (0.0279)*** (0.0420)** (0.0152)*** (0.00892)*** (0.00601)*** (0.0211)*** (0.0469)*** GARCH -Estimates Full Period the for (January 1979 April 1998) 1.007311 0.392083 -0.014719 0.243924 0.878390 0.248843 -0.423509 0.930958 0.694420 -0.131725 0.5319 0.999754 0.378161 -0.017124 0.263322 0.980901 (0.00520)*** (0.00576)*** (0.0647)*** (0.00296)** (0.0143)*** (0.00141)*** (0.00441)*** (0.0174)*** (0.00296) (0.0952)*** (0.00460) (0.0104)*** (0.00231) (0.00503)*** (0.0149)*** (0.00230) (0.00466)*** (0.0156)*** (0.00823)** (0.0305)*** (0.00449)*** (0.00410) (0.00966)*** (0.00613)*** (0.0364)*** (0.0281)*** (0.00249) (0.00265) 0.012837 0.012002 0.874084 0.286060 -0.024335 0.241152 0.001682 0.958155 Intercept R-ECU, 0.000496 0.739056 0.006701 1.003705 0.379582 -0.034522 (0.00275)** -0.001341 0.690207 0.138080 -0.029016 0.210922 0.971561 -0.003078 -0.012866 0.791643 0.375867 -0.103867 0.293926 0.896922 (0.00393)*** (0.00853)*** (0.0147)*** (0.0104)** (0.0244)*** (0.00658)*** (0.00402)*** (0.00765)*** (0.0226)*** (0.0164)** (0.0174)*** (0.00907)*** (0.00171)*** (0.00948)*** (0.0183)*** (0.0131)*** (0.0161)*** (0.00899)*** (0.00363)*** (0.00672)*** (0.0161)*** (0.00317)*** (0.00756)*** (0.0141)*** (0.0118)*** (0.0298)*** (0.00702)*** (0.00291)*** (0.00614)*** (0.0192)*** UK Italy 0.012309 0.817839 Spain France 0.002228 0.966204 Ireland Greece 0.038467 0.860204 0.230197 0.034913 0.975269 Sweden -0.002879 0.670793 0.249130 -0.527568 0.566678 Austria -0.008884 0.978841 Finland Portugal Country Belgium Germany Denmark Netherlands -0.003288 indicateslevel. significancethe1% at ** ** indicates significance the5% at level. * * *** indicateslevel, significancethe10% at Standard errors are in parentheses.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 97 0.4191 LM(IO)

Q.(io) 1.0451 1.0885 0.7291 0.7278 16.2457 16.2257 12.5463 12.4376 Table 1-1 B Panel Table 1-1 AIC LnL 4751.568 -2368.78 0.3483 4274.814 -2130.41 3067.127 3067.127 -1526.56 4.1875 4.8888 Diagnostics" of Full Time Period Time Full of Diagnostics" UK 5820.931 -2903.47 40.2473 41.8803 Italy 3963.53 -1974.78 13.3368 13.6046 Spain 3890.196 -1938.1 0.6781 0.7596 Greece 4217.191 -2106.6 0.4868 0.5039 France 2139.301 -1062.65 12.1734 12.4048 Ireland 6039.812 -3012.91 Sweden Austria Finland 4927.685 -2456.84 1.0624 1.1554 Country Belgium 2878.939 -1432.47 13.1374 13.919 Portugal Germany 2457.392 -1221.7 12.0522 12.0922 Denmark 3052.762 -1519.38 Netherlands 2455.495 -1220.75 (LM). (LM). Q(10) and denoteLM(10) the tests forthe significance ofresiduals to correlations the up in lag estimated standardized residuals, 10 e,/Vht. “ The diagnostics theare Akaike Criterion“ Information the (AIC), Log Likelihood portmanteau Q-test(LnL), the (Q), and multiplierLaGrange test

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 98 R' 0.5516 0.5006 0.6797 e (0.3064) (0.0561)** 0.390580 0.5398 0.145267 0.5277 0.261572 0.5376 0.382965 0.4804 0.590742 -0.322499 0.3296 -0.294993 0.3446 (0.0228)*** -0.148633 0.4171 Yi (0.0612) (0.0818)** 0.916956 0.604358 0.967348 0.378569 0.946958 0.969668 (0.0153)*** 0.960065 0.044237 -0.186128 0.4380 0.981033 (0.0703)*** (0.0597)*** 0.913041 -0.031991 0.852608 0.203879 (0.1064)*** (0.0248)*** 0095**(0.0669)*** (0.00915)*** ai 0.617902 0.823047 0.500464 -0.477562 0.5691 0.342953 (0.0455)*** (0.0187)*** (0.0700)**1 0.144745 (0.0736)*** (0.0421)*** (0.0427)*** (0.0369)*** (0.0739)*** (0.00990)*** (0.0383)*** (0.0844)*** 0) (0.0182)*** (0.0367)*** (0.0284)*** (0.0359)*** (0.0138)*** (0.0675)*** -0.055332 0.302980 0.939991 0.131971 0.3443 (0.0132)*** (0.0279)*** (0.00752)*** (0.0695)** -0.363822 1.142660 (0.0972)*** (0.0336)*** -0.042080 0.259100 (0.0142)*** (0.0368)*** (0.0627)*** (0.00845)*** (0.0147)*** -0.185572 1.272958 0.382504 0.136177 0.1948 (0.00934)** (0.0174)*** (0.0793)*** (0.00406)*** -0.041184 0.254133 (0.0367)*** -1.103599 0.698807 -0.021634 0.186833 -0.109080 0.484104 Table 1-2 R.-1 (0.0286) (0.1315)*** (0.0418)*** 0.425048 -0.045044 0.319043 (0.0329)*** (0.0610)*** 0.356740 -0.250284 0.500824 (0.0274)*** (0.0332)*** (0.0681)*** (0.0220)*** (0.0645) (0.00613)*** (0.0179)*** (0.0590)*** 10)*** 0.729581 0.299556 -0.512770 0.779861 0.453106 0.580883 0.346037 (0.0150)*** (0.0274)*** (0.0191)*** (0.0182)*** (0.0233)*** (0.0132)*** (0.0148)*** (0.0104) (0.0249)*** (0.0356)*** (0.0150) (0.0173)*** (0.00554) (0.0223)*** (0.0271)*** (0.00442) (0.0117)*** (0.0344)*** (0.00702) (0.0163)*** (0.0268)*** (0.00700) (0.00882) (0.00264)*** (0.0108)*** (0.0898)*** (0.00442) (0.0113)*** (0.0273)*** 0.024962 0.618319 0.022665 -0.841128 0.040303 0.771424 0.191468 0.006852 0.732718 0.425201 0.018941 0.663609 0.392693 -0.083160 0.024313 0.749383 -0.003339 0.581236 0.496622 -0.113167 0.000492 0.695096 0.273950 Intercept R-ECU, -0.021136 0.638376 0.173165 (0.00872)*** (0.0119)*** (0.000614** (0.0290)*** (0.000711)*** (0.00434)*** (0.01 (0.00544)*** (0.00557)*** (0.00173)*** (0.0170)*** GARCH Estimates forthe First Subperiod (January 1979 -December 1985) UK 0.006621 0.561582 0.142860 Italy Spain Greece Sweden France Ireland 0.008254 Finland -0.009215 Austria -0.004537 0.922507 Country Portugal Belgium 0.007593 0.764558 0.319378 Germany 0.005040 Denmark Netherlands ** ** *** indicates significance at the 5% level. indicates significance the at level. 1% * * indicates significance the at level 10% Standard errors are parentheses. in

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 99 R2 0.3057 9 (0.0503) (0.0926)** 1.194042 0.211324 0.4767 (0.0680)** (0.0479)** (0.0557)*** (0.0453)*** (0.0497)*** (0.0492)*** Yt 0.980284 0.218225 -0.107470 0.5137 0.916301 -0.119770 0.6376 0.457230 -0.247653 0.6130 0.999418 -0.606273 0.6193 0.987168 -0.151753 0.6423 0.538128 0.135348 0.6216 0.900378 -0.165840 0.6370 0.992190 0.087233 0.6427 0.666086 0.929562 -0.131446 0.6387 (0.0506)*** (0.0179)*** (0.0872)*** (0.0172)*** (0.0282)*** (0.00537)*** (0.00414)*** a t 0.190392 0.774074 0.778739 0.569371 0.076604 0.351455 0.962060 0.256134 0.5772 0.597754 0.191855 0.645880 0.413200 0.133602 0.571993 (0.0624)*** (0.0312)*** 0.771149 0.776095 -0.140945 0.5911 (0.0320)*** (0.0651)*** (0.0280)*** (0.00315)*** (0.0941) (O (0.0227) (0.1051)*** (0.0183)*** (0.0974)*** 0.008284 (0.00756)** (0.0298)* ♦♦ (0.00756)** (0.0298)* (0.0967)*** (0.0543)*** (0.0473)*** (0.0467)** (0.0505)*** (0.0292)*** (0.0553)*** (0.0166)*** (0.0448)** (0.0949)*** (0.0545)*** -0.220242 0.655587 0.821641 -0.007017 0.5631 (0.0312)*** (0.0347)*** (0.0468)*** (0.0238)*** (0.1563)*** -0.108395 -0.382658 (0.0456)*** (0.0653)*** (0.0273)*** (0.0711)*** -0.000052 -0.074058 -0.240646 Table 1-3 R,.i 0.259809 -1.385186 0.406823 -0.005035 0.415679 0.472384 (0.0331)*** (0.1513)*** (0.0598)*** (0.0331)*** 1.003074 0.417572 -0.089964 1.012332 0.447709 0.841348 0.408495 -0.719563 (0.0139)*** (0.0275)*** 0.977660 (0.0134)*** (0.00954)***II)*** (0.0252)*** (0.01 (0.00213)*** (0.1599)*** (0.00118)*** (0.00664)*** (0.0292)*** (0.000148)*** (0.00216)*** GARCH Estimates for the Second Subperiod (January 1986 -GARCH Estimates the for (January Subperiod Second 1986 December 1992) (0.00467) (0.00625) (0.0102)*** (0.0236)*** (0.00638) (0.0121)*** (0.0244)*** (0.00313) (0.00419)*** (0.00675)*** (0.00152) (0.00112)*** (0.0235)*** (0.00658) (0.00484) (0.00919)*** (0.00629)*** (0.00453) 0.005034 0.818867 0.189429 -0.015446 (0.00527)* 0.002591 0.910065 0.467397 -0.028450 (0.00754)* (0.0165)*** -0.017781 (0.00284)*** (0.0119)*** (0.00822)*** -0.004709 0.991052 -0.014082 (0.00763)*** (0.00198)*** (0.00525)*** (0.00540)*** (0.0275)*** (0.00123)*** (0.00554)*** (0.0180)*** (0.000251)*** UK Italy Spain -0.000949 0.885508 0.289382 Greece 0.040341 France -0.0015287 0.958416 Sweden -0.009068 0.738446 Ireland -0.009540 0.950988 0.422945 Finland 0.012865 0.721823 0.513866 Austria Country Intercept R-ECU, Portugal 0.006167 0.867336 0.304863 -0.900115 Belgium Germany Denmark -0.003951 0.948048 0.452016 Netherlands -0.008727 ** ** at the indicates significance5% level. * * *** the indicates at significancelevel 10% at the significance indicates level. 1% Standard errors are parentheses. in

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 100 R? 0.8094 0.7247 0.5477 0.4869 0.6352 0.8934 0.8372 e (0.0792) (0.1414)* 0.373790 0.4434 0.528012 0.6430 (0.1236)** 0.014386 (0.1263)*** -0.158187 0.4390 (0.1381)*** (0.0845)*** (0.0757)*** (0.1148)*** -0.651055 0.8875 -0.727234 0.8849 -0.051912 -0.268964 0.8347 Y> 0.987576 0.997451 0.987113 (0.00786)*** (0.00212)* *♦ (0.1074)*** (0.00212)* 0041**(0.1214)*** (0.00401)*** (0.00671)*** (0.00753)*** at 0.116380 0.996921 0.606845 0.087686 0.998396 0.279391 0.215455 0.982408 0.260491 0.128736 0.158526 (0.0234)*** (0.00623)*** (0.1283) 0.246172 (0.0198)*** 0.648453 0.741746 -0.327785 0.8383 (0.0131)*** (0.00233)*** (0.0513)*** (0.0392) (0.0173)*** (0.0284)*** (0.0551)*** (0.0315) (0.0544)*** (0.0466)*** 10 (0.00587) (0.0157)*** (0.0156)* (0.0406)*** (0.00740)*** (0.00713) (0.0213)*** (0.0324)*** (0.0699) (0.0120)** 0.003840 0.001958 -0.024564 0.104011-0.030124 0.982861 0.116601 0.982100 (0.0200)*** (0.0253)*** (0.00600)*** -0.070841 0.139987 0.975968 (0.0640)*** (0.0343)*** (0.0232)*** (0.1372)*** (0.0586)*** (0.0387)*** (0.0636)*** -0.345862 0.311348-1.830135 0.864861 0.614550 0.038558 0.473239 -0.535411 0.455512 0.843762 -0.200874 (0.1004)*** -0.045265 -0.044675 0.216274 0.977460 Table 1-4 R,., (0.0292) (0.0126)* 0.212214 0.143325 0.014328 -0.028536 0.194221 -0.031611 0.295410 0.087674 0.004525 0.224088 (0.0280)*** 0.229353 0.202362 -0.719035 -0.015387 (0.0275)*** (0.0137)** (0.0285)*** (0.00960)*** (0.00709) 1.008661 0.299567 1.052343 0.971265 0.181486 (0.0181)*** (0.0288)*** (0.0241)*** 0.982291 0.232273 (0.0105)*** (0.0154)*** (0.0240) (0.0137)*** (0.00845)*** (0.0878)*** (0.0154)*** (0.0296)*** (0.0102)*** (0.0234)*** (0.0203)** (0.00344)*** (0.0284)*** (0.00676)*** (0.00841)*** (0.0321)*** (0.0216)** GARCH Estimates- Third Subperiod (January the for 1993 April 1998) (0.0118) (0.00895) (0.00363) (0.00697) (0.00307) (0.00352) (0.0761)*** (0.00710) (0.00384) (0.00350) (0.00641)*** (0.0285)*** (0.00296) 0.007172 0.747946 0.006455 0.001863 1.052823 0.002736 0.737981 0.011507 0.824653 0.018942 0.965650 0.002995 0.980323 -0.010024 0.631931 0.092440 (0.00318)** (0.0103)*** 0.000556 1.096451 (0.00545)** -0.000894 0.994375 -0.001480 1.056381 (0.00475)*** (0.00460)*** (0.00990)*** (0.0288)*** UK Italy Spain 0.016970 France Ireland Greece Sweden Finland Austria Portugal Country Intercept R-ECU, Belgium Germany -0.001325 Denmark 0.000299 Netherlands * * ** *** indicates significancethe at level, 10% indicates significanceatthe 5% level. indicates significancethe at level. 1% Standard errorsStandard are parentheses. in

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 101 -44.76 -99.62 -94.21 -83.69 -81.94 -93.84 -89.44 -31.10 -51.80 -75.41 -99.93 -77.36 -70.94 -66.87 -43.99 0.194529 0.236493 0.2017220.0601432.594331 -23.34 -73.49 610.59 -12.46 -79.80 0.0597342.0214180.373696 -73.53 -60.86 -2.67 -79.70 -99.60 4.54 0.050897 0.069750 0.255547 0.086343 -63.54 -72.98 0.124050 Table 1-5 7.42 14.20 -98.99 -23.78 -27.96 -25.89 0.286174 -48.99 0.137928 0.3650972.238922 -99.95 -45.19 0.224847 0.312115 0.319543 0.236800 1st Period Avg.1st 2ndAvg.Period % Change2nd to 1st 3rdPeriodAvg. Change % 2nd to 3rd % Change 1st Averages of Daily Volatility per Subperiod and Percent Changes in Volatility between Subperiods between Volatility in Changes Percent and Subperiod per Volatility Daily of Averages UK 0.352173 0.282342 -19.83 Italy 0.230434 0.263153 Spain 0.561010 France Greece 508.978539 5.164587 Ireland 0.357450 0.383960 Sweden 4.084775 Finland 4.146587 382.107220 9114.98 Austria 0.221480 0.374485 69.08 Portugal 677.661197 Country Belgium Germany 0.294256 0.225701 -23.30 Denmark 0.386206 0.240027 -37.85 Netherlands 0.297681 0.226900

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

102

Wall StreetJournal Wall Table 2-1 Table by the Bundesbank. the by issuing short-term debt. short-term issuing considered. Swiss put faith in the Euro to drive down the value of the Swiss franc, to reduce current strengthening. current reduce to franc, Swiss the of value the down drive to Euro the in faith put Swiss introduced. from ERM. from A poll of European companies shows support for a single currency. single a for support shows companies European of poll A U.S begins to trade futures on the ECU. the on futures trade to begins U.S Private use of the ECU is made legal. made is ECU the of use Private Summary of Announcements Obtained from the the from Obtained Announcements ofSummary Positive Positive The 15 governments agreed upon a new name for the single currency and set 1999 as the date the currency is to be to is currency the date the as 1999 set and Positive currency single the for name new a upon agreed governments 15 The Positive Three-step process toward implementing a single currency is set forth. Positive set is currency single a implementing toward process Three-step Positive Positive Negative Realigning ERM hurts EMS and slows progress toward a single currency. single Negativea toward progress slows and EMS hurts ERM Realigning 1/8/90 1/8/86 8/3/88 PositiveECUs. in denominated debt long-term issues U.K. 1/21/97 market. single a and currency single a to commitment Positivetheir underscore governments German and French The 5/6/88established. and Positive discussed are currency single a maintain and develop to Conditions 10/2/95 5/29/97objections of spite in EMU, to due reserves gold of Positiverevaluation the for presses Kohl, Helmut Germany, of Chancellor 1/11/95 removed be to peseta Spanish the cause Negative would which crisis currency of beginning like looked what is Witnessed 3/12/97be implementation in delay one-year a that Negative suggests 1999, by ready be not will governments many that Recognizes 6/14/96 uncharacteristically by specifically currency, Positivesingle a of inception the suit to policy monetary changes Germany 10/4/88 Positive ECUs. in denominated debt short-term issues U.K. 5/30/95 fashion. proper a in currency single a Positive achieving to commitment Underscores currency. single for timetable new Sets 4/15/96 Positive negotiated. are stable currency new the keep to Methods 5/17/905/13/91 Positive Positive ministers. finance the by overcome are currency single a to Obstacles 8/28/89 Unit. Currency .European the for Positivewitnessed is optimism more removed are barriers trade As 2/20/92 approved. are PositiveECUs in denominated payments clearing for Arrangements 11/18/96 12/18/95 6/17/87 2/23/88 Positive bank. central joint a for push a Announces 12/24/92unveiled. Positive are crisis currency a following ECU's of use the reviving for Plans 9 8 6 7 5 1 4 19 2 3 17 18 22 15 16 21 20 14 12 13 11 10 Event# Date Positive/Negative Summary

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

103 R**2 0.5237 0.4278 0.5879 e (0.0306)*** (0.0365)*** (0.0266)*** -0.131725 0.5319 -0.208680 0.6265 Yi 0.969404 0.362646 0.694420 0.870499 0.081354 (0.0189)*** (0.00295)*** (0.00702)*** (0.0394)*** 1.0885 0.5039 LM(I0)

QdO) 0.241152 0.930958 0.330296 0.958258 0.236416 0.975269 1.0451 0.4868 (0.0210)*** a ai LnL -2106.6 -1221.7 12.0522 12.0922 (0.0164)** (0.0174)*** (0.00907)*** (0.0104)** (0.0244)*** (0.00658)*** (0.0427)*** -1974.78 13.3368 13.6046 0.034913 -2130.41 (0.00583)** -0.024335 -0.423509 -0.052304 -0.014719 0.243924 0.9836UO Table 2-2, Panel B Panel 2-2, Table Table 2-2, Panel A AIC R,-, 3890.2 -1938.1 0.6781 0.7596 3963.53 4274.81 4217.19 2457.39 0.293971 0.392083 (0.0161)*** (0.0356)*** (0.0347)*** (0.0152)*** Diagnostics of AR(1) - AR(1) Results ofEGARCH(1,1) Diagnostics Italy Spain Greece Country Portugal Germany R-ECU, 1.007311 0.874084 0.286060 0.878390 0.248843 0.817839 (0.00672)*** (0.00756)*** (0.0141)*** (0.0118)*** (0.0298)***

• P

AR(1)ARCH(1,1) - EG Estimates for Period (January 1979- April 1998) r*l O O (0.00249) (0.00520)*** 0.012002 0.012309 0.038467 0.860204 0.230197 0.012837 O, -0.003078 (0.00402)*** (0.00765)*** (0.0226)*** (0.00393)*** (0.00853)*** (0.0147)*** (0.00363)*** Italy Spain Greece Country Intercept Portugal Germany and LM(10) denote the tests for the significance of residuals correlations up to lag 10 in the estimated standardized residuals, SiA/ht. residuals, standardized estimated the in 10 lag to up correlations residuals of significance the for tests the denote LM(10) and * The diagnostics are the Akaike Information Criterion (AIC), the Log Likelihood (LnL), portmanteau Q-test (Q), and the LaGrange multiplier test (LM). (LM). Q(10) test multiplier LaGrange the and (Q), Q-test portmanteau (LnL), Likelihood Log the (AIC), Criterion Information Akaike the are diagnostics The * *** *** level. 1% the at significance indicates * * ** level. 10% the at level. significance 5% the at indicates significance indicates Standard errors are in parentheses. in are errors Standard

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 104 — — — — — — — — ~ — — — — — — — — 9.31% 81.22% 50.66% 28.16% 196.55% 0.109921004 0.188843459 0.064153616 0.089431680 39.40% 0.061745406 0.076496982 23.89% 0.0843473500.115135917 -33.70% 0.223041295 0.089815199 0.446278442 0.239234834 0.094768128 0.127215937 0.270113846 0.2590543870.273513755 -4.09% 0.348217389 0.239657180 19.63% 0.577751205 0.295194628 0.537796791 0.2279441210.200330409 -57.62% 0.194822596 ~ — — — — — — — — — — — — — — — 50.04% 0.062901504 -42.78% 94.87% 0.077054318 -24.93% 26.90% -11.75% -22.96% 0.061748222 -46.37% -20.71% 0.106444857 -52.28% -70.61% 0.132331003 -51.62% 3250.22% 0.132421950 0.140051030 0.162693144 0.157103426 0.138643885 0.067954867 0.102639376 1.290766893 0.185875934 0.143204861 0.231691080 0.183706490 0.053368127 0.078843852 47.74% 0.162761201 0.126535539 0.151998392 -56.75% 0.429099566 0.4680131S8 0.137557000 0.412735701 0.605344663 0.418820953 -67.55% 0.325544973 31.39% 0.247770491 0.424861665 13.827555561 10.864012300 -65.13% 0.360436630 31.155452110 — — — — — — — — — — — — — — 4.62% 5.32% -0.25% -7.50% 0.202444654 24.43% 0.206421316 50.22% -48.80% 0.114677893 192.35% -18.57% -42.34% Italy Change % Greece Change % Spain Change % 0.0804268270.041179407 0.076434079 0.137803748 0.363021089 0.524347804 0.485008912 0.141473467 0.148546231 0.131827009 0.351439000 0.073159818 0.0555182110.266123059 -24.11% 0.257638598 0.033768808 0.090895261 169.17% 0.152364207 20.41% 0.129813903 36.98% 0.212664283 0.433384094 0.187955532 0.263572564 0.197913019 0.563557860 0.190627995 0.175455821 Table 2-3 — — — — — ...... _ — — _ — — — _ — — ... Change 3.88% 19.60% 0.115208050 2.40% 60.26% 0.454702505 4.92% -34.94% 0.182823232 -14.03% 0.544523144 -42.99% 0.207649950 8.93% % % 0.0817450020.069856116 -12.65%0.072566200 0.144167186 0.112930610 0.093588351 0.133047066 0.071737248 0.150868827 13.40% 0.374493836 3.16% 0.154975665 10.66% 0.1217133590.295366322 59.96% 0.168838602 0.362558435 0.172280030 -52.48% 0.778012817 0.359507689 0.355096237 0.076089668 0.163214107 -27.13% 0.116563090 -37.98% 0.380225668 0.223984938 0.306161933 0.314619090 0.433348279 0.170736067 -45.73% 0.180265458 -68.01% 0.231104056 -61.82% 0.204388203 — — ~ — — — — — — — — _ — — — — — Change Portugal 88.72% 36.45% 0.269559015 -25.02% 0.271341849 -27.36% -48.99% 0.154757804 21.54% -11.04% -57.43% -24.42% 0.159497032 -47.90% 0.139855878 -29.33% 0.272774842 -35.80% 0.179990204 -39.03% 56 56 Percent Changes in Average Daily Volatility One Month prior to and after Announcements 0.105485384 0.127332785 0.014548243 0.355478863 0.055477778 0.186074773 -23.03% 0.247034778 0.283456619 0.290800036 0.177771064 0.326971855 14.91% 0.209301995 To Germany 1/9/95 8/1/88 1/17/96 0.086584590 2.30% 0.159646824 41.37% 0.148173688 6/10/92 0.031189926 -43.78% 0.141859590 -51.97% 0.076007971 5/15/90 0.029473756 8/30/88 0.383303263 31.81% 0.247361380 1/5/90 2/5/90 0.485065681 From 1/7/86 2/5/86 5/5/88 6/3/88 8/2/88 7/1/88 1/17/92 2/18/92 0.036023530 12/8/97 1/10/95 2/8/95 0.027491311 88.97% 0.085798051 9/1/88 9/30/88 0.248557278 4/6/88 5/4/88 0.241747433 6/13/96 7/12/96 0.022081658 8/30/95 9/28/95 0.092857090 5/26/95 6/26/95 0.053809188 3/15/96 4/12/96 0.048470345 9/29/95 10/30/95 0.200143774 115.54% 10/3/88 11/1/88 0.139279076 -43.96% 4/15/96 5/13/96 0.030217388 -37.66% 1/21/88 2/19/88 0.665892791 5/10/91 5/16/90 6/15/90 0.055622683 2/19/92 3/18/92 0.038518581 6.93% 0.089266212 -47.13% 4/27/95 5/25/95 12/5/85 1/6/86 0.284557051 8/25/89 9/25/89 0.586103053 34.56% 0.569088067 4/11/91 5/9/91 4/17/90 7/27/89 8/24/89 0.435558796 11/15/95 12/14/95 0.084636153 12/15/95 6/16/87 7/14/87 0.134366209 2/22/87 3/21/88 11/23/92 12/22/92 0.054677084 12/23/92 1/25/93 0.048640825 8 12/5/89 1/4/90 9 1 5 6 7 17 18 5/14/96 6/12/96 0.030397344 4 15 16 23 5/15/87 6/15/87 11 12 13 14 10 fentU

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 105 — — — — 88.32% 0.045779567 3.71% 0.079902859 0.0488667490.044140390 -38.84% 0.041173169 0.077538536 0.059712405 0.096869133 62.23% — — -- -- -4.53% -49.06% 0.106329438 0.101517306 0.284143710 0.144732653 0.128022295 0.115410754 -9.85% 0.123898249 0.114920766 -7.25% — — — — 79.19% 0.081725910 -4.05% 0.0830385770.085173993 -59.66% 0.0751486880.077899033 -2.13% 0.139590843 0.205845679 0.076786105 -- — — — 0.135813152 58.84% 0.0762962780.085505408 -27.34% 0.075910006 0.0873738710.105002742 15.10% 0.294483462 0.132918267 -54.86% — — — — 13.06% 96.49% 0.044378332 89.66% 0.053092498 0.0372243030.023398855 -29.89% 0.049172547 0.044190469 0.049960737 0.025025639 4/8/97 1/16/97 5/27/97 3/10/97 12/16/96 1/17/97 2/18/97 5/28/97 6/25/97 3/11/97 11/15/96 10/16/96 11/14/96 19 21 2/7/97 22 4/28/97 20 12/17/96

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

106

Wall StreetJournalWall Summary Table 3-1 Table Summary of Announcements Obtained from the the from Obtained Announcements of Summary Recognizes that many governments will not be ready by 1999, suggests that a one-year delay in implementation be implementation in delay one-year a that suggests 1999, by by ready be objections of not will spite in EMU, to governments due many reserves gold that of Recognizes revaluation for the presses Kohl, considered. Helmut Germany, of Chancellor Bundesbank. the Swiss put faith in the Euro to drive down the value of the Swiss franc, to reduce current strengthening. current market. reduce to single a and franc, Swiss the of currency value single the a to down drive to commitment Euro their the in faith underscore put Swiss governments German and French The Germany changes monetary policy to suit the inception of a single currency, specifically by uncharacteristically issuing uncharacteristically by specifically introduced. currency, single a of inception the suit to policy monetary changes Germany debt. short-term Methods to keep the new currency stable are negotiated. are stable currency new the keep to Methods Witnessed is what looked like beginning of currency crisis which would cause the Spanish peseta to be removed from removed be to peseta Spanish the cause would which crisis currency of beginning like looked what is Witnessed Sets new timetable for single currency. Underscores commitment to achieving a single currency in a proper fashion. proper a in currency single a achieving to commitment Underscores currency. single for timetable new Sets The 15 governments agreed upon a new name for the single currency and set 1999 as the date the currency is to be to is currency the date the as 1999 set and currency single the for name new a upon agreed governments 15 The ERM. Three-step process toward implementing a single currency is set forth. set is currency single a implementing toward process Three-step Positive Positive Positive Negative Negative Date Positive/Negative 5/29/97 Positive 10/2/95 Positive 11/18/96 Positive # 9 3/12/97 8 1/21/97 1 1/11/95 7 5 4/15/95 6 6/14/96 10 3 4 12/18/95 Positive 2 5/30/95 Positive Event

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 107 Rf 0.0018 e (0.1854) (0.1280) (0.1164) (0.1200)*** -0.003410 0.1561 -0.004901 0.0485 -0.671427 0.0518 n 0.985546 0.958580 -0.221031 0.0320 0.975234 -0.158642 0.0411 0.936258 (0.0481)*** (0.5273) -0.859310 0.815048 0.0149 -0.792359 -1.460622 0.0300 -0.192172 0.078952 (0.00617)*** CCl (0.0294)*

0.165482 0.209672 0.149110 (0.0400)*** (0.0134)*** (0.1421) -0.056893 -0.421912 10 (0.0165)** 1.414164 0.055751 0.002779 0.072413 0.990551 012)* (0.0464)***(0.1724)*** (0.0916)** Table 3-2, Panel A R,; 0.189456 0.165012 002)* 0050* (0.0436)*** (0.0113)*** (0.00530)** (0.0326)*** (0.0350)*** -0.040406 0.039548 0.251517 (0.0711) (0.0359)*** (0.0145)*** (0.0347)*** (0.0174)*** 0.125965 -0.053993 -0.083435 0.075790 GARCH Estimates forthe Country ADRPortfolios (0.0588) (0.0316) (0.0474)*** (0.0298) (0.0170)** (0.0583)*** (0.0189)*** (0.0757)*** (0.0408) (0.0561)*** (0.0458) (0.0544)*** (0.0311) (0.1052)*** (0.0395) (0.0927)*** (1.3034) 0.080107 0.041152 (0.0317)*** (0.0352)*** (0.0292)*(0.0118)*** (0.000734)*** (0.0811) (0.0298)* (0.00201) (0.0197)*** (0.0312)*** (0.0273)*** (0.0115)*** (2) (5) (4) (1) (1) (1) (1) (17) (0.0290) (0.0375)*** UK -0.012721 0.310471 0.152050 0.010915 France 0.028481 0.318814 0.017099 Ireland Greece -0.065593 0.316750 0.213010 0.038096 Sweden 0.048045 0.482579 -0.056258 Finland 0.088887 -0.074685 0.031030 1.874024 Country Intercept R-NASDAQ, Netherlands Luxembourg -0.000817 0.231593 *** *** indicates significance at the level. 1% * * ** indicates significancethe at level. 10% indicates significancethe at 5% level. Standard Standard errors are parentheses. in NumberofADRs each in country portfolio underthe country is parenthesis. name in

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 108 8,/Vh,. 5.6151 11.4668 LM(10) QdO) 5.5601 5.3104 9.3655 8.7505 9.1733 9.6908

Table 3-2, Panel B 3650.289 3650.289 -1818.14 4.8201 6.3463 3135.531 3135.531 -1560.77 3745.475 3745.475 -1865.74 Diagnostics1of theADR Portfolios UK Greece 3988.929 -1987.46 9.8725 10.657 France Ireland 5107.349 -2546.67 5.1935 Sweden 3021.085 -1503.54 12.3187 Finland 4698.958 -2342.48 6.2495 6.2571 Country AIC LnL Netherlands 2892.344 -1439.17 Luxembourg 1 The1 diagnostics are the Akaike (LM).Information Q(10) Criterion and (AIC),LM(10) the denote Log Likelihood the tests for(LnL), the significanceportmanteau of Q-testresiduals (Q), correlations and the LaGrange up to multiplierlag 10 in the test estimated standardized residuals,

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 109 — — — — — — — — — — Change 44.21% -23.83% % % 8.440331003 22.00% 7.829567842 8.735403132 -40.73% 6.918333551 7.171147996 3.716043207 7.626966975 4.549642136 4.841587775 14.738434830 10.005436210 18.421476410 — — — — — — — — — — Change Ireland 1.26% 5.963893568 14.69% 5.358928003 21.98% 54.89% 85.97% 7.649285330 0.29% -30.85% 15.041948550 -18.35% % % Greece 1.917830018 3.192209241 3.152584333 1.095701373 2.897103732 4.487187324 4.728987930 4.104506464 4.707660207 4.988150847 3.067822507 4.410891957 3.524316957 0.757641104 — — — — — — — — — — Change 0.22% 3.566587669 -4.62% 5.768512644 -0.68% % % Table 3-3 2.269282741 4.80% 2.165344908 2.287177456 2.164288665 2.2229071642.240985809 2.71% 3.122759345 -37.40% 8.875352788 -11.29% 2.225841832 2.2182238102.206626062 2.14%2.202472049 3.610498748 -0.19% -18.15% 6.942587169 -3.19% 2.197671322 2.153144886 -2.03%2.171797171 3.581249069 16.74% 6.297999333 38.43% 2.193008289 2.199115027 2.237607726 — — — — — — — — — — Change France -4.73% % % 6.044741910 6.86% 5.656803389 5.796937127 -5.00% 2.181528077 5.951274131 -2.14% 5.891898522 6.184446233 5.881613262 -2.48% 2.252792774 0.68% 2.750047470 -21.97% 5.529378093 14.21% 6.297855963 5.35% 2.197934724 2/8/95 5.775201440 -2.85% 2.245119062 2.09% 5/13/96 9/28/95 5.978022612 11/14/96 6.045516887 Percent Changes in Average Daily Volatility One Month prior to and after Announcements 1/17/97 2/18/97 5/28/97 6/25/97 3/11/97 4/8/97 12/8/94 1/9/95 5.944719767 1/10/95 6/13/96 7/12/96 5.725175366 4.48% 5/26/95 6/26/95 9/29/95 10/30/95 4/15/96 11/15/96 12/16/96 6.109671545 1.06% 12/15/95 1/17/96 5.876803890 -2.71% 11/15/95 12/14/95 6.040621811 8 12/17/96 1/16/97 6.081346571 7 10/16/96 9 2/7/97 3/10/97 6.102063893 1 5 3/15/96 4/12/96 6 5/14/96 6/12/96 5.479850530 3 8/30/95 4 10 4/28/97 5/27/97 2 4/27/95 5/25/95 6.031011961 ent #ent From To Finland

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. HO % % % % % % % % % % 02 24 05 99 . 72 80 82 . 48 . — . 06 — — — — 60 . — . — . — — — . . . 5 Change 2 3 17 2 34 - - 15 63 - - - - % % UK 1.08922413 1.116278066 3.83516557 1.313120686 1.051907019 1.724997228 1.881465637 1.392787774 1.182926827 8 1.757454283 0.926245486 1.782666206 2.221014139 18 0.907998236 1.539497702 1.130377069 0.934377411 1.107085029 2.672538759 3.973383352 % % % % % % % % % % 20 10 92 00 28 05 . . 60 . . 37 . — . — — 68 — 25 — — . — — — — . . . Change 1 5 12 26 14 22 21 - 29 - - - - % % Sweden 1.20721662 1.56619334 0.66165244 1.135406061 14 1.206097945 1.370585255 0.995104529 0.572311965 0.754594939 0.836120342 1.301582104 0.908718518 8 1.004303786 0.775539929 0.775539929 1.006435598 0.765971449 0.952349351 0.742255225 0.942962222 % % % % % % % % % % 33 62 12 44 57 38 — — — — 03 82 . — — . . — 91 . — — — 04 ...... 0 0 2 Change 0 0 0 1 0 ------% % 0.96639777 0.96691263 0.964705131 0.953426653 0 0.967941013 0.964682411 0.970718228 0.942843961 0.953123765 0.973833703 0.965932272 0.964722518 0.970267431 0.961318449 0.979701462 1 0.968524941 0.969660354 0.965370415 0.975454874 0.970608776 Table 3-3, continued % % % % % % % % % % 84 82 40 18 38 31 84 . . . . 38 . 28 . 76 . — — — — — . — — . — — — . 7 16 11 32 42 35 32 23 - 27 46 ------1.404194850 1.421247199 1.750657678 1.849107694 1.630220070 1.285227318 1.947318064 1.540239235 1.296304617 1.017658136 1.649224552 1.107653976 2.230703908 Luxembourg % Change Netherlands 0.840880058 0.911665280 3.127927101 2.138342856 16/96 1.619717747 14/96 8/97 14/95 30/95 8/95 18/97 10/97 2.077263140 9/95 12/96 16/97 12/96 13/96 17/96 12/96 2.380971977 27/97 25/97 / / / 28/95 25/95 / / / / / / / / / / / / / / / / 4 1 1 2 1 5 6 5 5 (,126195 12 9 11 12 10 Percent Changes in Average Daily Volatility One Month prior to and after Announcements 8/94 15/96 17/96 16/96 7/97 3 15/95 15/95 11/97 17/97 2 13/96 7 14/96 / 15/96 15/96 4 28/97 28/97 6 10/95 / / / / 29/95 26/95 / 30/95 27/95 / / / / / / / / / / / / / / 2 From To 1 5 3 4 5 1 12 6 4 3 8 5 9 11 12 10 12 8 7 9 6 5 1 4 11 3 10 2 4 Event #

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 111 CHARLOTTE ANNE BOND

College of Business Administration 3211 Nobscot Drive, Apt. C Butler University Indianapolis, Indiana 46222 Indianapolis, Indiana 46208 (317) 925-2147 (317)940-8154 [email protected] [email protected]

AREAS OF INTEREST: Teaching: International Finance, Corporate Finance, Investments

Research: International Financial Issues, European Monetary Issues, Investments

POSITION DESIRED: Assistant Professor with balanced teaching and research responsibilities in graduate and undergraduate programs.

ANTICIPATED AVAILABILITY: August 1999

REGIONAL PREFERENCE: None

EXPERIENCE: Butler University Visiting Assistant Professor 1998 Full-time Old Dominion University Research and Teaching Assist. 1995-98 Part-time Lamar University Instructor of Economics 1994-95 Part-time Lamar University Research Assistant 1993-94 Part-time

Taught courses in Corporate Finance, International Financial Management, and Principles of Microeconomics and Macroeconomics

EDUCATION: Old Dominion University Finance and 1995-98 Ph.D. International Business

Lamar University Business 1993-94 M.B.A. Georgia Institute of Technology Management 1988-1993 B.S.

HONORS: Beta Gamma Sigma Lamar University 1994 Dean’s List (6 terms) Georgia Tech 1991 -1993

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 112 DISSERTATION: “Three Essays on European Monetary Union Advances toward a Single Currency and Implications for Business and Investors”

PUBLICATIONS: “Structural Models of Exchange Rate Determination.” co-authored with Mohammad Najand, submitted to The Journal o f Multinational Financial Management

SUBMISSIONS: “Volatility Changes in European Currency Exchange Rates.” co-authored with Mohammad Najand, to Global Finance Journal.

“Changes in European Currency Volatility as Related to Changes Occurring during Europe 1992.” co-authored with Mohammad Najand, toThe Journal o f International Money and Finance.

“Volatility Changes in European American Depository Receipts Returns: Evidence from the NASDAQ Market.” co-authored with Mohammad Najand, to European Economic Review.

“European Equity Market Integration.” co-authored with Mohammad Najand, submitted to The Journal o f International Money and Finance.

PRESENTATIONS: “Volatility Changes in European Currency Exchange Rates.” co-authored with Mohammad Najand, presented at the 1998 Financial Management Association Conference.

“European Equity Market Integration.” co-authored with Mohammad Najand, presented at the 1997 Financial Management Association Conference and the 1997 European Financial Management Association Conference

“Structural Models of Exchange Rate Determination.” co-authored with Mohammad Najand, presented at the 1997 Eastern Finance Association Conference

“European Monetary Union: Implications for Business.” presented at the 1996 Academy of International Business U.S. Northeast Regional Conference

“Dynamics of the International Automotive Market: Can the American Auto Industry Thrive in the Global Market?” co-authored with Mark Fincher, presented at the 1996 Academy of International Business U.S. Northeast Regional Conference

OTHER SCHOLARLY ACTIVITIES: Assistant to the Vice President - Arrangements, 1998 Eastern Finance Association Conference, Williamsburg, Virginia

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Participant in the 1998 Konrad Adenauer Stiftung (Foundation) German-American Seminar “The Double Challenge: European Integration and Globalization” which included discussions with economists, diplomats, politicians, business people and social workers regarding the opportunities and threats of globalization and integration for European Union countries.

Discussant at the 1997 Financial Management Association Conference

REFERENCES: Mohammad Najand Charles Hawkins Sylvia Hudgins Department of Finance Dept, of Economics and Finance Department of Finance College of Business Lamar University College of Business Old Dominion University Beaumont, Texas 77710 Old Dominion University Norfolk, Virginia 23529 (409) 880-8647 Norfolk, Virginia 23529 (757) 683-3509 (757) 683-3551

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. IMAGE EVALUATION TEST TARGET (Q A -3 )

■*U450 1.0 ■ us ISi | 2.2 !f 2.0 l.l 1.8

1.25 1.4 1.6

150mm

irVWCBE . Inc 1653 East Main Street Rochester, NY 14609 USA Phone: 716/482-0300 Fax: 716/288-5989

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.